Back
[00:01:04] <alex4nder> tjb1: is there room for 8 chips?
[00:01:12] <tjb1> There is room for 5
[00:01:35] <tjb1> 2 on each end of the front side and a place in the middle for one chip
[00:01:43] <tjb1> nothing on the back of the stick
[00:02:01] <alex4nder> dunno
[00:02:15] <alex4nder> why are you trying to track that down?
[00:02:20] <tjb1> I need to get more ram
[00:02:48] <alex4nder> oh, and you're trying to find something compatible?
[00:02:52] <tjb1> yes
[00:03:32] <tjb1> All it says on dell spec sheet is "PC2700 or PC3200 DDR SDRAM (non-ECC)"
[00:03:46] <alex4nder> then just get whatever
[00:03:47] <tjb1> 128, 256, 512, 1 sticks
[00:03:49] <alex4nder> unbuffered
[00:03:55] <alex4nder> non-ECC
[00:04:29] <tjb1> I do believe its low density, things I've read say 64Mx64 is low and I have 4 chips so 64/4 = 16Mx64
[00:07:01] <tjb1> Guess its too hard to throw that marking on the chip with the 40 other numbers on there...
[01:11:10] <Loetmichel> mornin'
[01:51:26] <Jymmm> hola
[01:51:45] <alex4nder> sup Jymmm
[01:51:56] <Jymmm> fire
[01:52:28] <Jymmm> lots and lots of fire (hopefully)
[02:07:37] <DJ9DJ> moin
[03:47:41] <alex_joni> yu Jymmm
[03:47:46] <alex_joni> yo* even
[03:52:30] <skorket> Does anyone have experience with grbl?
[03:55:26] <Jymmm> alex_joni: ug
[03:55:54] <Jymmm> alex_joni: i got a hawt date with a pillow so make it quick =)
[03:56:46] <alex_joni> heh
[03:56:50] <alex_joni> I'll /msg you
[04:34:06] <Jymmm> alex_joni: you know anything about thermo sealing plastics?
[05:25:27] <alex_joni> Jymmm: not really
[05:48:49] <alex_joni> should we port LinuxCNC to the Shiri?
[05:49:24] <alex_joni> http://www.youtube.com/watch?v=vhHo6CUq4-o&feature=player_embedded
[08:03:48] <skunkworks> that is just creepy
[09:21:08] <joe9> i have a (MSG, depth is [#<depth>])
[09:21:19] <joe9> i am trying to print the actual depth value.
[09:21:40] <joe9> the message I get is "depth is [#<depth>]"
[09:22:00] <joe9> is there any way to get it to print the actual value held by the variable #<depth>
[09:23:53] <joe9> got it "In the DEBUG, PRINT and LOG comments, the values of parameters in the message are expanded."
[09:25:51] <alex_joni> try without []
[09:26:09] <alex_joni> and isn't it #_<depth> ?
[09:26:23] <joe9> alex_joni: just as a fyi, it worked with DEBUG.
[09:26:43] <joe9> the manual says that MSG is not expanded with special operators
[09:27:12] <alex_joni> ok then :)
[09:50:53] <joe9> i am trying to make some clearance holes for a #10 screw. From the docs, I can read that I need a #7 drill for a loose fit. I am wondering if I should just use a #7 drill or use an end-mill to make a hole of that size.
[09:52:07] <skunkworks> joe9: I use end mills all the time to make clearance holes.. also counter sinking
[09:52:24] <skunkworks> sometimes it is easier than finding a drill...
[09:53:29] <joe9> skunkworks: ok, thanks.
[09:56:18] <JT-Shop> my chart says #10 screw uses a #10 drill for a clearance hole
[09:56:27] <jdhNC> that woudl be too convenient.
[09:56:44] <JT-Shop> it's the only one that is that way
[09:57:22] <jdhNC> mine says tight #7, loose C
[09:57:39] <jdhNC> for a 10-32
[09:59:56] <JT-Shop> my chart must be tight
[10:00:35] <jdhNC> another says close 9, free 7
[10:01:40] <JT-Shop> only 0.0025 difference in diameter between 9 and 10
[10:02:48] <jdhNC> mine came from my $1. android app, then the web.
[10:05:34] <skunkworks> I ususally use 3/16...
[10:05:34] <JT-Shop> mine came from the giant Kennedy chart haning on the wall...
[10:09:38] <joe9> http://www.stanford.edu/~jwodin/holes.html i got the clearance hole size from here.
[10:10:40] <skunkworks> saw mill still works...
http://electronicsam.com/images/SawMill.JPG
[10:10:47] <joe9> skunkworks: i am looking to buy some endmills. any recommendations on the common ones to have around?
[10:11:14] <joe9> i am thinking of
http://www.mcmaster.com/#8977A183
[10:11:21] <skunkworks> 1/4 carbide... 1/2 carbide...
[10:11:30] <joe9> http://www.mcmaster.com/#8977A153
[10:11:35] <joe9> http://www.mcmaster.com/#8977A123
[10:11:38] <skunkworks> I usually order from enco for most things,,,
[10:12:17] <skunkworks> You can get them a lot cheaper
[10:12:18] <joe9> 1/4 and 1/2 are the end mill diameters, correct?
[10:12:26] <joe9> ok, thanks. will check out enco.
[10:14:27] <skunkworks> yes
[10:16:09] <joe9> skunkworks: do you have any urls of the end mills you order from enco?
[10:16:28] <skunkworks> joe9: usually whatever is on sale :)
[10:16:34] <joe9> oh, ok. thanks.
[11:24:52] <joe9> is this line wrong? G42 G01 X#<x1> F#<feed_rate> D#<tool>
[11:25:03] <joe9> i notice that it is not turning on the radius compensation.
[11:25:14] <joe9> not sure if I am missing something simple in that line.
[11:28:17] <joe9> is there any way I can print the x value which is used when G42 is in force?
[11:29:58] <archivist> are you reading the docs properly
[11:30:23] <archivist> dont mix gcodes
[11:37:13] <Aero-Tec> hi
[11:38:45] <Aero-Tec> is there a way to display the values of variables while running Gcode?
[11:39:42] <Aero-Tec> would like a running display, not something you have to hit the OK button to get rid of each time
[11:40:37] <Aero-Tec> so no pop ups unless the popup was updated and did not block working with the GUI of EMC
[11:49:03] <joe9> Aero-Tec: check out PRINT
[11:49:30] <joe9> Aero-Tec: no, that needs the ok button
[11:50:20] <joe9> i am using this to make a rectangle slot.
http://codepad.org/mIkZAl3l
[11:52:05] <archivist> joe9, a codepad with a code pad address...wtf
[11:52:21] <joe9> I notice that when the cutter is at x1 = 0.3, actual x value = 0.2375; y1 = 1.0, actual y value = 0.9375; x = 0.7, actual x value = 0.7625 and when y = 2.0, actual y value = 2.0625
[11:52:30] <joe9> archivist: sorry about that. wrong pasted.
[11:52:37] <archivist> what makes you think line 5 is valid g code
[11:52:55] <joe9> http://codepad.org/rGYWb2RN
[11:53:27] <joe9> archivist: good one. let me double check on that. thanks.
[11:53:43] <joe9> there were no errors, so I figured it must be ok.
[11:53:51] <joe9> let me change that and try.
[11:54:40] <Jymmm> hi andypugh
[11:54:46] <joe9> archivist: is there any way I can get a grid or something with more values in the backplot.
[11:55:26] <andypugh> joe9: A grid in the backplot was added to the dev branch a day or two ago
[11:55:29] <archivist> some are experimenting with a grid, not in the released axis
[11:55:40] <joe9> i have the big line with the start and end points of the axes.
[11:55:44] <joe9> andypugh: cool, thanks.
[11:56:53] <archivist> joe9, dou you want g0 or g1, why are you typing g01
[11:57:13] <joe9> archivist: i want g1
[11:57:26] <joe9> let me change that to G1.
[11:57:53] <andypugh> G01 seems to be used quite frequently, but I have never been clear why
[11:58:04] <archivist> why is your D1 in the middle of nowhere
[12:00:03] <joe9> archivist: D1 is the parameter to G42
[12:00:44] <archivist> I realise but after the g1!!!!
[12:01:12] <archivist> use the syntax in the manual
[12:01:22] <joe9> i changed the lines to
http://codepad.org/bvzTlN2W and still it has the same offsets.
[12:02:57] <joe9> should I be switching between G42 and G41 for different lines.
[12:03:10] <IchGuckLive> hi all
[12:03:35] <jdhNC> is there a problem with your offsets? It looks like you have a .0625 radius tool?
[12:04:30] <joe9> jdhNC: yes, that is what I have.
[12:04:40] <joe9> i am trying to draw a rectangular slot.
[12:04:53] <jdhNC> with cutter radius compensation
[12:04:56] <joe9> this is the code
http://codepad.org/bvzTlN2W
[12:05:04] <joe9> yes, with cutter radius compensation.
[12:05:13] <joe9> I notice that when the cutter is at x1 = 0.3, actual x value = 0.2375; y1 = 1.0, actual y value = 0.9375; x = 0.7, actual x value = 0.7625 and when y = 2.0, actual y value = 2.0625
[12:05:20] <IchGuckLive> joe9: use heekscad
[12:05:29] <joe9> could not get it to install.
[12:05:39] <IchGuckLive> there is a postprocessor for this
[12:05:59] <IchGuckLive> joe9 witch system
[12:06:14] <joe9> IchGuckLive: linux amd processor.
[12:06:24] <IchGuckLive> 10.04
[12:06:47] <joe9> jdhNC: i probably need a G42 followed by a G41 and then a G40 to get it work correctly.
[12:06:54] <joe9> jdhNC: would you agree?
[12:06:59] <jdhNC> so you tell it to go to X0.3Y1.0 and it puts the edge of the cutter there.
[12:07:25] <IchGuckLive> joe9 do you got a dxf
[12:08:03] <IchGuckLive> or give me the edge coordinates lower left upper right
[12:08:20] <IchGuckLive> inside or outside climb or conventionell
[12:08:29] <joe9> IchGuckLive: my gcode :
http://codepad.org/v0ztkjAU
[12:08:41] <jdhNC> from what I see, it is doing exactly what you told it to do.
[12:09:57] <IchGuckLive> joe9 did you got a look on our simple g-code generators on the wiki
[12:10:14] <joe9> i am trying to draw a rectangular slot of (0.3,1) (0.7,2). but, with the code there I get a slot of (0.3,1) (0.8ish,2.12)
[12:10:31] <IchGuckLive> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Simple_LinuxCNC_G-Code_Generators#Rectangular_Circular_Pocketing_Generator
[12:10:42] <joe9> i probably need a G41 too in the code.
[12:11:11] <IchGuckLive> the G42 G41 depends on inside climb outside conventionell
[12:11:40] <joe9> IchGuckLive: i tried that, but, could not get the "Write to Axis and Quit" to work.
[12:11:47] <joe9> and the "To Clipboard" did not work.
[12:11:54] <joe9> there was nothing being copied to the clipboard.
[12:12:08] <IchGuckLive> i wrote this script
[12:12:25] <IchGuckLive> witch version of axis are you running
[12:12:47] <joe9> LINUXCNC - 2.6.0~pre
[12:13:14] <joe9> how do i find the axis version?
[12:13:21] <joe9> IchGuckLive: it just says "quit" for me.
[12:13:31] <joe9> it does not have the "Write to Axis and quit".
[12:14:37] <joe9> and, it does not copy to clipboard. is there a debug version that I can try with more messages?
[12:15:00] <joe9> btw, why the gui. why not just a cli which will output the program to stdout?
[12:15:11] <Aero-Tec> so there is no running display of variables in EMC?
[12:15:55] <Aero-Tec> I am making springs and I need to know what the number of springs made is
[12:16:01] <IchGuckLive> joe9 i checked it running perfect here
[12:16:15] <IchGuckLive> witch linux distro you got
[12:16:27] <joe9> i use a distro called crux.
[12:16:34] <IchGuckLive> my git is from tonight at linuxcnc
[12:17:10] <joe9> let me upgrade linuxcnc then. mine is atleast a month old.
[12:17:29] <IchGuckLive> thats not the case
[12:17:34] <andypugh> Aero-Tec: Which variables?
[12:17:42] <Aero-Tec> how does one monitor variables when trouble shooting gcode?
[12:17:50] <Aero-Tec> any
[12:18:07] <Aero-Tec> just need a counter to say how many springs are made
[12:18:19] <Aero-Tec> ant counter would work
[12:18:23] <Aero-Tec> any
[12:18:28] <andypugh> Permanently or temporarily?
[12:18:39] <Aero-Tec> temp for this code
[12:18:58] <Aero-Tec> perm if necessary
[12:19:09] <Aero-Tec> but temp would be better
[12:19:34] <Aero-Tec> I am using the mill to make springs
[12:19:43] <andypugh> M66 (I think) will set motion.analog-out-00, you can view that with Halscope, or longer-term make a GladeVCP panel to view it.
[12:19:58] <Aero-Tec> so it is just a temp setup when I make springs
[12:20:43] <Aero-Tec> how many digits?
[12:20:51] <Aero-Tec> need 4
[12:21:20] <Aero-Tec> 3 could work if I brake it into 2 runs
[12:21:21] <djdelorie> 12V 50A $50 :
http://www.newegg.com/Product/Product.aspx?Item=N82E16817341017
[12:21:43] <Aero-Tec> 2 would be a real pain
[12:22:27] <Aero-Tec> could one setup a phantom axis?
[12:22:55] <Aero-Tec> have a axis display that had no motor and move it to increase the display?
[12:23:23] <jdhNC> DJ: what does one do with a high amp 12v PSU?
[12:23:30] <andypugh> Aero-Tec: Halmeter shows 6 or more digits
[12:23:33] <Aero-Tec> how many axis will the GUI display?
[12:23:35] <djdelorie> someone was talking about it yesterday
[12:23:44] <djdelorie> i.e. the need for high-amp 12v supplies
[12:23:58] <Aero-Tec> ok will look into it
[12:24:02] <Aero-Tec> thanks
[12:24:25] <Aero-Tec> is there other outputs one can use at the same time?
[12:25:18] <Aero-Tec> like motion.analog-out-01, 02, 03?
[12:26:10] <joe9> IchGuckLive: could you paste the code that is generated with the parameters: rectangle, parallel, 0.3,1 -> 0.7,2
[12:26:20] <joe9> IchGuckLive: if you do not mind.
[12:26:35] <joe9> IchGuckLive: i am upgrading linuxcnc now to the latest git release.
[12:26:40] <IchGuckLive> i will
[12:26:44] <joe9> thanks.
[12:26:51] <IchGuckLive> cutter size
[12:27:14] <joe9> radius = 1/16 inch.
[12:27:29] <andypugh> djdelorie: Better still, have a look on eBay for Fruit Machine power supplies, typically 50A @ 35V, 20A @ 12V, 5V, 24V...
[12:27:46] <IchGuckLive> im in mm it takes 5min for putting all togetter in inch
[12:27:47] <djdelorie> it all depends on what your steppers want
[12:28:07] <djdelorie> my servos want 160v at around 8 amps, for example, but they're not getting it :-)
[12:28:11] <andypugh> Of course, I can't find them now..
[12:28:14] <joe9> IchGuckLive: oh, ok. what ever is easy for you.
[12:28:30] <djdelorie> but if you have 3.3v steppers, a 12/50 supply might be just right.
[12:28:38] <joe9> IchGuckLive: put the radius as 1mm and 3mm,1mm -> 70mm, 20mm.
[12:28:50] <joe9> IchGuckLive: i want to read/understand the code generated.
[12:28:50] <andypugh> djdelorie: That sounds easy if you are in the US. Just rectify 110V mains, straight into a cap.
[12:29:04] <djdelorie> I thought of that. I want *some* ground isolation :-)
[12:29:19] <djdelorie> I suspect the motors were designed for exactly that, with isolation on the logic side of the controller
[12:30:04] <djdelorie> an isolated CAN interface is trivial, that would be a likely candidate. Or a constant-RPM use. etc.
[12:32:09] <alex4nder> oddly enough, I built an isolated CAN transceiver carrier board last night
[12:32:33] <alex4nder> automotive is a bitch
[12:34:01] <andypugh> djdelorie: Just use optos between the PC and the drives?
[12:34:22] <andypugh> Or, run through an isolation transformer
[12:34:44] <djdelorie> I considered that, but my controllers have not only the pport interface, but also two UARTs and a CAN bus. Isolating them all would have been prohibitively expensive
[12:35:00] <djdelorie> and I wouldn't be able to debug the board with my usual logic analyzer
[12:35:28] <djdelorie> I considered isolation between the MCU and the driver, but there's five analog feeback circuits there too
[12:35:34] <ReadError_> awwww yea
[12:35:36] <Jymmm> djdelorie: "usual logic analyzer" == LED + resistor?
[12:35:40] <ReadError_> cut3d made a good toolpath i think
[12:35:59] <djdelorie> no, a $400 32-channel 500 MHz real logic analyzer
[12:36:28] <Jymmm> djdelorie: That's an expensive resistor, I hope you dont need a 4000W one =)
[12:36:37] <alex4nder> laf.
[12:36:59] <jdhNC> Read: what did you use for input to cut3d?
[12:37:09] <Jymmm> djdelorie:
http://www.reopowersolutions.com/product_list/28
[12:37:47] <alex4nder> djdelorie: TI is sampling their ISO1050 isolated CAN transceiver.
[12:38:31] <djdelorie> I don't use braking resistors. The linear supply is designed for 2x the working voltage, and has excessive caps. I use regenerative braking :-)
[12:38:56] <andypugh> I have a 1:1 building-site transformer that I use for isolation when I am fiddling, but I am not that bothered about it when everything is assembled.
[12:39:17] <djdelorie> doing the isolation is easy, but the parts cost money and take up board space, whereas a simple linear supply solves all those problems and protects me when I'm debugging
[12:39:31] <Jymmm> Man, a $500 resistor that's just nuts. And the worlds worse waster of energy! lol
[12:39:38] <djdelorie> but a light bulb!
[12:39:59] <Jymmm> djdelorie: at least with a light bulb you get light from it
[12:40:00] <ReadError_> jdhNC: STL
[12:40:14] <djdelorie> every cnc machine needs a 4000 watt blinky light :-)
[12:40:56] <Jymmm> djdelorie: But a resistor, it just sits there, doens't do anything, doesn't look cool or pretty, just a pure energy whore!
[12:41:51] <alex4nder> it'll heat your home
[12:42:04] <Jymmm> alex4nder: not very good though
[12:42:15] <Jymmm> compared to gas
[12:42:56] <Jymmm> though It seems to work for iron/steel mills
[12:43:50] <Jymmm> Man, if they could recover the radiant heat and "store" or distributed it somehow
[12:44:14] <IchGuckLive> xyuv howto is up ->
http://www.youtube.com/watch?v=ZYHpkH_xHVA
[12:44:25] <IchGuckLive> HD full screen
[12:45:01] <gene__> Andy- Silly pid Q. Pgain,Igain,Dgain=0.0. FF0=1 command=301, output all over the map. With all those gains at 0, where is the dither coming from
[12:45:07] <gene__> Andy- Silly pid Q. Pgain,Igain,Dgain=0.0. FF0=1 command=301, output all over the map. With all those gains at 0, where is the dither coming from?
[12:46:19] <joe9> /ws 33
[12:48:26] <Aero-Tec> how hard is it to add axies?
[12:48:36] <Aero-Tec> I need more axis on my mill
[12:48:45] <Aero-Tec> hardware is done
[12:49:07] <Aero-Tec> but EMC step wis will only help with setting up 4
[12:49:24] <Aero-Tec> I need 5 with more coming
[12:50:01] <Aero-Tec> and will the GUI see them and display them?
[12:50:42] <Tom_itx> just add it in the hal and ini files
[12:50:44] <Aero-Tec> I am converting over from Mach
[12:51:07] <archivist> I just edited the hal and ini when I went up to 5 axis
[12:51:13] <andypugh> The output is varying with no change to input?
[12:51:22] <andypugh> Is the error value changing?
[12:53:59] <IchGuckLive> Aero-Tec: then just add it manuell
[12:54:29] <IchGuckLive> in the ini the Axis number XYZABCUVW x=0
[12:54:43] <IchGuckLive> in HAL ad one more stepgen
[12:55:04] <IchGuckLive> its mutch easyer then you expect
[12:55:43] <andypugh> gene__: Is the error value steady?
[12:55:53] <Aero-Tec> ok so it is easy to figure out
[12:56:00] <Aero-Tec> thanks
[12:56:41] <Aero-Tec> so would the files I am looking for be .hal and .ini?
[12:57:11] <IchGuckLive> yes there is also a advise to clone from git to the newest linuxcnc to get all the nice features we all made
[12:57:37] <IchGuckLive> Aero-Tec: yes in the mashine name folder under config
[13:02:28] <IchGuckLive> Aero-Tec: GeometryAXES = 3 this number has to be 1 higher then the AXIS Geometry Axis value
[13:03:29] <IchGuckLive> so example XYZAB is your config you need here the C axis number 5
[13:04:02] <IchGuckLive> Aero-Tec: what is your geometry
[13:04:17] <IchGuckLive> witch axis are you running XY???
[13:04:51] <Aero-Tec> will have to look
[13:05:16] <IchGuckLive> not good to know the axis at the start of the config
[13:05:36] <IchGuckLive> 5Axis can also mean XYBUV
[13:05:58] <Aero-Tec> that do you mean by "what is your geometry"?
[13:06:20] <IchGuckLive> as your nick seams to get a foamcutter as here for aeroprofiling
[13:06:25] <Aero-Tec> xyzab
[13:06:52] <Aero-Tec> no fome
[13:06:54] <IchGuckLive> thats so easy
[13:07:02] <Aero-Tec> foam
[13:07:21] <Aero-Tec> mills and lathes
[13:07:57] <IchGuckLive> just copy the stepconf wizard A axis [AXIS_3] and rename it to [AXIS_4]
[13:08:23] <IchGuckLive> same in hal and cghange the stepgen number one higherv
[13:09:25] <IchGuckLive> [TRAJ] section AXES=5 COORDINATES = X Y Z A B Home add a zero
[13:09:39] <IchGuckLive> done in ini
[13:10:25] <Aero-Tec> will try it
[13:10:28] <Aero-Tec> thanks
[13:11:44] <IchGuckLive> in hal loadrt stepgen step_type=0,0,0,0,0
[13:12:59] <IchGuckLive> stepgen.4 then tot the pins you will see this
[13:13:47] <IchGuckLive> if you got -> net Apos-cmd axis.3.motor-pos-cmd => stepgen.3.position-cmd
[13:14:11] <IchGuckLive> add below net Bpos-cmd axis.4.motor-pos-cmd => stepgen.4.position-cmd
[13:14:28] <IchGuckLive> Axis 4 is your B
[13:14:47] <IchGuckLive> stepgen 4 holds all the waveforms
[13:15:17] <IchGuckLive> just walk throu the file
[13:17:48] <IchGuckLive> im missing pcw is he on holyday
[13:22:16] <IchGuckLive> ok im off thanks for the evening nice talk B)
[13:22:44] <IchGuckLive> Thunderstorm hedding my way no grilling alowed
[13:22:55] <IchGuckLive> O.O
[13:26:25] <gene__> Andy, no
[13:28:44] <gene__> timing the index pulses, at a command, and output of pid at 101, I am seeing about 114 rpm
[13:30:38] <gene__> Andy, that does change by much when I bring up some Pgain
[13:34:37] <gene__> The A & B pulses positions wobble by only 10% of each pulses width, but encoder.0.velocity is seeing 40% wobbles in halscope
[13:35:55] <gene__> the A/B wobble is half a rev from the index using delayed sweep
[13:44:23] <gene__> 200 ms after the index, the A pulse is consistent to within 1 ms. I can only come to the conclusion that the encoder is inventing these 30% speed variations.
[13:46:35] <gene__> I just urned pwmgen dither off, no real change
[13:46:36] <archivist> are you using a real scope or halscope
[13:48:35] <gene__> both
[13:50:01] <gene__> the rea scope is measuring the index period to determine rpms, and it delay to measure timing errors of the A pulse train anywhere up to 200ms after the index
[13:50:34] <archivist> can you hear the speed changing, have you added a mass to the spindle to damp it a bit for testing
[13:51:10] <archivist> are the opto signals clean
[13:52:29] <gene__> yes I can hear the hum of the motor changing, and a 5" 4 jaw chuck is the main mass, several pounds, the A signal is straight out of the opto's on a real 100 mhz dual trace scope
[13:54:07] <gene__> pwmgen dither is off, its output is steady at about 15% duty cycle at this speed
[13:55:36] <gene__> Under these conditions, no feedback IOW, the motor hum & rotation are smooth
[13:59:42] <gene__> Interesting. My optos on the encoder board are not bolted down, and while its difficult to get a finger in there, I found I can reduce the error p-p by pulling outward on the front one to increase the separation. The quadrature looks good on the big scope, so I'll cross check it.
[14:01:22] <archivist> home brew encoder?
[14:02:58] <gene__> yup, next Q, which edge of the index does it use?
[14:04:40] <archivist> are the edges clean on the signal with NO noise
[14:10:36] <gene__> absolutely
[14:11:48] <gene__> BRB, coffee cup on missing list
[14:12:23] <andypugh> gene__: Sorry, I was away.
[14:13:42] <andypugh> It's very difficult to test, but on my homebrew encoder the quadrature offset drifts a bit around the disc. It means that unless the mark-space ratio of both enocders is right, then some pulses are not counted
[14:13:49] <gene__> np Andy
[14:14:47] <gene__> its very close tp 50%
[14:17:11] <gene__> I do see that the A pulse must be seeing part of the index slot, 1 wide pulse per rev, by aout 10 wide
[14:17:18] <gene__> I do see that the A pulse must be seeing part of the index slot, 1 wide pulse per rev, by aout 10% wide
[14:17:57] <andypugh> Ah, the index isn't just a longer version of another slot?
[14:18:51] <andypugh> It might be interesting to set counter-mode, that will answer the question if it is quadrature error.
[14:23:06] <gene__> yes and no, it extends inward from slot 0 but was cut seperately. The 3 optos are in a row, straight line, so the center one rides the index & the outer 2 furnish the A & B from the slot pattern
[14:24:54] <gene__> counter-mode=1, error gone
[14:25:17] <jdhNC> do a&b read different slots?
[14:25:37] <gene__> yes, but in the same circle
[14:27:42] <archivist> is the index too wide
[14:27:53] <archivist> relative to the normal slots
[14:30:05] <gene__> not quite the same width, pulse is about 10% wider because the gcode that carved it made the slot side face the centerline, so the outer end of the index slot is a thou, maybe 2, wider than the inner root of the main slots
[14:31:12] <gene__> Not enough room, and to move the center opto up off the board would mean I'd have to file off the outside edge of the wheel
[14:31:26] <gene__> off for clearance
[14:32:58] <gene__> Dopes running it this mode mean it only does the error update on the index pulse? That is what I am seeing in halscope
[14:33:14] <gene__> Does running it this mode mean it only does the error update on the index pulse? That is what I am seeing in halscope.
[14:34:27] <gene__> I see the magnitude of pid.0.feedback is scaled down by about a /4 too
[14:39:29] <gene__> I setp that scale to 240, setp Pgain to 5, wants to oscillate on a speed change, but 3 works
[14:44:55] <gene__> bringing up the Igain make the error go positive
[14:47:56] <gene__> set that scale down to 225 seems to balance that
[14:58:57] <andypugh> Yes, in counter-mode you only see 1/4 the pulses, so the encoder scale needs to change.
[14:59:17] <andypugh> You also lose the ability to detect direction, I don't know if that matters.
[14:59:43] <andypugh> when you say "error update" what do you mean?
[15:00:04] <gene__> pid.error
[15:00:17] <andypugh> That is updated every servo-thread
[15:00:26] <gene__> ahh, encoder scale needs to be 39 then
[15:01:47] <DJ9DJ> gn8
[15:06:16] <gene__> Ok, if I wanted to ditch the scale.1 thingy, can I set the encoder scale to .65, and get that gain of 60 back but without the time delay?
[15:08:15] <andypugh> Eh?
[15:12:44] <andypugh> If you don't care about the onscreen speed display, you don't need the scale at all.
[15:14:02] <andypugh> You can set encoder.scale to 39, and then that gives velocity in rps. Then use the speed-command-rps signal from motion so that both command and feedback are the same. You can then either re-scale the pwmgen to an rps full-scale, or just re-tune the PID to 60x bigger numbers
[15:21:27] <Solaris> I got lotsamoney now
[15:21:29] <Solaris> I want a CNC machine
[15:21:39] <Solaris> Shall I buy the X2 and a gecko?
[15:21:44] <Solaris> http://www.ebay.co.uk/itm/Axminster-SIEG-X2-SuperTilt-Head-Mill-Drill-/270974353230?pt=LH_DefaultDomain_3&hash=item3f17547f4e
[15:24:59] <andypugh> Nah, you want:
http://www.ebay.com/itm/120898634649
[15:24:59] <andypugh> \
[15:26:09] <Solaris> i was thinking something in the 2000GBP price range
[15:28:17] <andypugh> Ah, OK.
[15:28:27] <andypugh> That might get you a ready-converted one.
[15:28:41] <Solaris> UK based, I can't import from America
[15:28:42] <andypugh> Heck, it would probably buy mine off of me, if I was finished with it.
[15:28:58] <Solaris> Too much customs and tax
[15:29:07] <andypugh> It's in Basildon
[15:29:25] <gene__> Well, I just found I can't set the encoder to output rpms. drives g76 out of its mind ;) So much for what seemed like a good idea :(
[15:30:24] <andypugh> Read back..
[15:30:52] <andypugh> Leave the encoder in rps, use the rps output from motoin as the command.
[15:30:59] <andypugh> Solaris:
http://www.amadeal.co.uk/acatalog/copy_of_Syil_X2_CNC.html
[15:31:06] <andypugh> Possibly a little smaller than you want
[15:32:26] <andypugh> gene__: YOu get that?
[15:32:44] <andypugh> (I do hope so, this is the 4th time I have suggested it in the last 24 hours)
[15:35:02] <andypugh> Solaris: You are not the only one in the UK. I am, archivist is, several others too.
[15:35:42] <archivist> should I put my backside as my google image :)
[15:37:16] <andypugh> To show it is too narrow to be American, you mean?
[15:37:27] <archivist> but thinking of second image down
http://www.bbc.co.uk/search/news/?q=vintage%20computer
[15:37:35] <jdhNC> that's a lot of money for an x2
[15:38:07] <andypugh> Yeah, it is pre-converted though. It's easy to lose track of how much a conversion costs
[15:38:11] <archivist> andypugh, I felt slim on holiday over there :)
[15:38:38] <jdhNC> I keep trying to lose track of how much I have spent on mine.
[15:38:47] <archivist> and I was a british standard fat git at the time
[15:39:44] <Connor> Me Too! :)
[15:40:11] <Solaris> Surely all I need for a basic system is an X2 mill, 3 steppers, a gecko g550 a suitable powersupply some wire and a pc?
[15:40:26] <Connor> Honestly, I've not spent a ton.. Most went to the motors and electronics.. I've probably got $100-150 in wire, connectors and misc crap..
[15:40:42] <djdelorie> I think I've spent more on developing those controller cards than on everything else.
[15:40:43] <Connor> Solaris: Couplers
[15:40:47] <archivist> Solaris, they change the leadscrews too
[15:41:17] <Connor> Stock to make the stepper mounts.. hardware to mount them..
[15:41:39] <Connor> Misc tools to make them.. taps etc ..
[15:41:54] <jdhNC> I ended up using about 24" of Al round stock and that's it.
[15:41:59] <archivist> I spent dunno how much at the scrap yard on bits of metal for making mounts etc
[15:42:04] <jdhNC> still have all th plate
[15:42:13] <Solaris> Is the coupler to tap the rotor of the steppers onto the leadscrew?
[15:42:14] <Connor> jdhNC: Not true, you used some plastic you had for the mounts..
[15:42:41] <jdhNC> right, but I already had that for something else. $30 ft^2 delrin
[15:43:06] <Connor> every $ adds up.. If you didn't have it, you would have bought it...
[15:43:10] <archivist> Solaris, the stock manual leadscrew has too much play(end float/backlash)
[15:43:21] <jdhNC> I went to the junkyard, found tons of perfect huge stainless stock. No good mild or Al
[15:43:36] <ReadError_> hmmm
[15:43:39] <ReadError_> visualmill is pretty nice
[15:44:06] <jdhNC> ball screws and ACB's added $250
[15:44:46] <Solaris> So I want a trapezoidal thread and a ballscrew, can I use the existing X2 rails?
[15:45:30] <jdhNC> Solaris: read through Hoss's X2 stuff
[15:47:34] <andypugh> trepzoidal thread and ballscrew are two completely different things. And ballscrew is better.
[15:48:04] <andypugh> Solaris: Cheapest ballscrews in the UK are probably from Zapp, www.zappautomation.co.uk
[15:48:24] <andypugh> You might also want to look at his stepper drives, as Geckos get very expensive if you import from the US
[15:48:27] <jdhNC> andypugh: how much is cheap?
[15:48:58] <andypugh> jdhNC: Look at the site, but about $1 / in ?
[15:49:42] <jdhNC> I paid $199us shipped for my chinese ballscrews/nuts machined
[15:50:10] <archivist> we get taxed rather a lot here
[15:50:14] <andypugh> Oooh! They have 6mm ballscrews now, how cute!
[15:52:27] <Connor> V belt or Timing belt for convert my spindle from gear to belt driven ?
[15:52:39] <andypugh> Why? Quietness?
[15:52:54] <Connor> speed and quietness
[15:53:32] <andypugh> You don't need to provide a tensioning mechanism for timing belt, if you get the centres right.
[15:53:32] <archivist> I use car v belts
[15:59:38] <Connor> So, after making a new encoder disk last night that has a index slot, and using index only.. I was able to do threading... worked good.. except I don't have the correct bit for it.. but, it hit the mark every time! :)
[16:00:25] <jdhNC> grind one
[16:00:51] <jdhNC> what were you using?
[16:01:05] <Connor> All I have to grind one with is my little benchtop grinding and it's wheel has a ding in it.
[16:03:24] <Connor> http://www.shars.com/products/gallery/3398/404-2048Main.jpg/
[16:03:38] <Connor> from that set, the closest one.
[16:04:21] <djdelorie> my set looks like this:
http://www.delorie.com/photos/southbend-lathe/img_2627.html
[16:04:34] <Connor> djdelorie: ROFL
[16:04:36] <andypugh> One of those is trying to be a threading tool
[16:04:46] <gene__> Andy - took time for some lunch now trying to do hal again as you asked.
[16:04:48] <Connor> andypugh: Which one?
[16:04:54] <andypugh> The pointy one
[16:05:08] <andypugh> but djdelorie has a more _useful_ set of tools
[16:05:13] <djdelorie> anything with a 60 degree tip could be used as a threading tool
[16:05:26] <djdelorie> andypugh: it's not useful if I have no idea what they're all for
[16:05:27] <Connor> That the closest one? That's the one I was using..
[16:05:35] <andypugh> High-speed steel is perfectly good enough for anything other than industrial mass production
[16:05:56] <djdelorie> but at least I KNOW that grinding my own tools is part of the job... the existing tools make that really obvious :-)
[16:06:30] <djdelorie> some of the bigger ones appear to be ground for mouldings, no less...
[16:07:31] <andypugh> djdelorie: I see a flycutter, an internal thread chaser (i think) a travelling steady, a parting tool, a boring bar holder, a couple of QC toolpost holders, and lots of useful bits of HSS
[16:08:49] <djdelorie> I've figured out the QC stuff, yes. The flycutter isn't that useful as there's no easy way to mount work to the sled. There are a bazillion HSS bits :-)
[16:09:02] <andypugh> The big bit in the middle of 2627 bolts ot the holes in 2628, but I assume from their numerical propinquity that you had figured that out?
[16:09:30] <djdelorie> yes
[16:09:46] <gene__> Andy - now I have lost the speed Plus and Minus buttons.
[16:09:51] <djdelorie> I need to clean out those holes and figure out the thread, and I'm missing bits on the following rest anyway. I did pick up a steady rest though
[16:10:23] <djdelorie> (that's *why* I took those two pictures, BTW ;)
[16:10:35] <Solaris> Is anyone here familliar with "lost wax" casting?
[16:10:51] <Solaris> http://i.imgur.com/DyZ52.jpg
[16:10:58] <Solaris> I made a latex negative of my gear
[16:11:07] <andypugh> gene__: Really? That's AXIS being too clever. It has noticed that you are not using motion.speed-command. Just net dummsig motion.spindle-command
[16:11:12] <Solaris> I tried to plaster of paris the backside of the latex but it got stuck in
[16:11:27] <Solaris> Is there another way I can make use of the latex negative and stop it wonking out of shape?
[16:11:29] <djdelorie> squishy gears!
[16:11:51] <andypugh> Solaris: I am vaguely familiar with it, but I have never done it
[16:11:53] <Solaris> The latex is so when the wax is poured I can peel the mold off, as they are spur not bevel gears sand doesn't work, I tried
[16:11:58] <Connor> djdelorie: What's with the long bits of spiral ?
[16:12:23] * JT-Shop has to make a 1 TPI double lead thread
[16:12:26] <djdelorie> they're shavings, proof that I'm cutting correctly :-)
[16:12:43] <andypugh> Solaris: What are the gears for?
[16:12:47] <djdelorie> Solaris: sand?
[16:13:04] <djdelorie> 2669 is a single eight-foot long shaving of aluminum
[16:13:21] <JT-Shop> iirc I just have to back up my Z start point by 1/2" right
[16:13:39] <djdelorie> I picked up some 7x7/8 bolts dirt cheap at the flea market, they turn MUCH easier than the hardware store steel rod stock. I have tiny steel curlies now too :-)
[16:14:25] <gene__> motion.speed-command doesn't exist, lemme check speeeling
[16:15:59] <gene__> No help there?
[16:16:10] <Connor> http://www.ivdc.com/cnc/conversion/encoder_disk1.JPG
[16:16:13] <Connor> http://www.ivdc.com/cnc/conversion/encoder_disk2.JPG
[16:16:36] <Connor> New encoder disk next to old one.. First truly automated part my CNC made..
[16:16:42] * djdelorie needs to make a 60-slot encoder to measure my lathe speed...
[16:17:37] <Connor> djdelorie: OKay, what's with the water and battery charger setup about ?
[16:17:48] <djdelorie> electrolysis rust removal
[16:18:00] <djdelorie> see
http://www.delorie.com/photos/southbend-lathe/img_2544.html
[16:18:09] <djdelorie> before on right, after on left
[16:18:42] <Connor> Anything special needed ? Can't tell how you have it hooked up..
[16:19:08] <djdelorie> 1 tablespoon washing soda (sodium carbonate) per gallon, steel anode to positive, part to negative. That's it
[16:19:27] <Connor> You using the battery, or the battery charger ?
[16:19:33] <djdelorie> add a battery or capacitor in parallel to avoid any 0.7 volt "wrong way" voltage for cheap one-diode chargers
[16:20:22] <djdelorie> I had a battery handy, so the charger charges the battery (pointless, it's a dead battery, but it does the job well enough) and I use jumper cables for the hookup so I don't mess up my charger clips
[16:20:35] <djdelorie> all you really need is to run an amp or two through the part
[16:21:48] <djdelorie> I'm just following what I read online, it's a popular enough topic among antique restorers.
[16:22:45] <gene__> Andy, thats a non-starter, any way I spell it
[16:23:14] <gene__> motion.spindle-command does not exist
[16:26:00] <gene__> I have .spindle-speed-in and .spindle-speed-out, which is it?
[16:28:04] <Solaris> I want to make loads of big gears so I can make my own recycled sky tv dish field that follows the sun and makes clean water and electricity
[16:28:28] <Solaris> Its to make lots of these:
https://www.youtube.com/watch?v=BAWLjqMiulo
[16:28:50] <djdelorie> one of my "one of these days" projects is to put an extruder head on my cnc mill and see if it can make 26" wide gears...
[16:29:50] <Solaris> Oooh yeah
[16:30:12] <Solaris> Apache helicopter gears
[16:30:16] <djdelorie> *in theory* I can do up to 26" diam by 3.5" thick. There are a lot of assumptions behind that.
[16:30:51] * djdelorie suspects shrinkage during cooling will be a huge pain
[16:31:01] <Solaris> No powercut, no overheat, no "out of oil" no stepping lsoses, etc.
[16:31:26] <djdelorie> and helical is just as easy as straight :-)
[16:31:46] <Solaris> With my lost wax method I learned you can get around shrinkage by having four copes at the bottom at the lowest point and for drags at the top where the spout hole is
[16:31:55] <Solaris> The its the copes and drags that shrink, not the part you want
[16:32:38] <djdelorie> with extrusion, the problem is that new layers stick to the old cold layers, then shrink, so the whole part bows up a little and messes it all up
[16:32:56] <djdelorie> you have to design in shrinkage reliefs into the extrusion pattern
[16:35:01] <gene__> Ok, working again, but scales are afu. speed increments show as rps, do as way off rpms.
[16:35:48] <gene__> and its about 50 clicks to top speed of 1100, shows 200 at motion.rps
[16:36:29] <Connor> okay, so, does anyone know how to get EMC to recognize the onboard parport for the intel atom boards? I've been using a addon PCI parport card.. but, I can see the need for using the onboard one at somepoint.
[16:36:50] <djdelorie> if they have different addresses, you just put in the two addresses.
[16:37:07] <djdelorie> if they're usb-based pport emulators, you're out of luck.
[16:37:33] <Connor> Yea. I can do that with the dual-port card.. but, I can't see the other one.. It's built in pport, so, I don't think it's a USB emulator.
[16:37:40] <gene__> atom has the std address of 0x378
[16:38:04] <djdelorie> make sure your extra card isn't the *same* address, and make sure your built-in is enabled in BIOS
[16:38:39] <Connor> It's not, addon is 0x1008 and 0x1018 or something similar..
[16:38:53] <djdelorie> so just put in 0x378 and see if it works
[16:38:58] <cei> Whenever I try and access the wiki, I get a 403 Forbidden error. Am I unique in this?
[16:40:43] <djdelorie> what exact url ?
[16:41:10] <cei> http://wiki.linuxcnc.org or /any/ link to it
[16:41:34] <djdelorie> 403 for me too
[16:41:54] <andypugh> gene__: spindle-speed-out
[16:41:54] <djdelorie> wiki is broken
[16:42:40] <andypugh> broken here too
[16:43:05] <cei> nice to know its not just me :)
[16:43:55] <gene__> as in "net dumbsig spindle-speed-out"
[16:44:00] <Keknom> same here
[16:44:27] <Keknom> also is machining 7050, and 7075 aluminum much diffirent from 2024?
[16:45:56] <gene__> Ok, did that, large rpms error, with an FF0=60
[16:47:32] <gene__> s250 gives 4.166667 in rps, gets 2.64
[16:48:42] <gene__> http://www.linuxcnc.org/docs/devel/html/config/emc2hal.html is working fine here
[17:04:30] <andypugh> gene__: Did you change the pwm scale, or not?
[17:05:13] <andypugh> Anyway, some value of FF0 should make it work OK..
[17:08:01] <gene__> pwm.scale=1100 in rpms
[17:10:04] <gene__> I think my feedback is scaled wrong atm, so i made a printout, easier to trace. with encoder.scale=slots (39), then encoder.velocity is rps, right?
[17:11:56] <andypugh> I think so. If you are in counter mode. Did that work better, or not?
[17:13:21] <gene__> for a few, till I tried to commit it all to the .hal
[17:20:39] <gene__> s300, spindle-rps=5, ah scale.1.gain for feedback was still 60, 1.0 looks a lot better
[17:21:18] <gene__> And I can get rid of the time lag of that scale then, brb
[17:28:13] <gene__> Done, speeds off some
[17:29:39] <gene__> pid.0.command=5 pid.0.feedback=5.4 on average pid.0.output s/b 300 is 298 average
[17:30:16] <gene__> pid.error=-.4 average
[17:31:47] <gene__> and I'm getting about 5.44 rps average from the index pulse spacing
[17:34:02] <andypugh> Sounds good, put some P and I in, and it should sort itself out
[17:34:08] <gene__> so I'll clean off the end of my rod & try another thread
[17:41:13] <andypugh> A bit of a rarity:
http://www.ebay.co.uk/itm/290705864050
[18:01:40] <gene__> fucking touch off on mthe wrong axis just ate me again, damn this is getting old
[18:02:04] <JT-Shop> did you make the pyvcp touch off buttons?
[18:03:05] <gene__> nop, haven't seen the code
[18:03:09] <JT-Shop> stop making me look at flea bay before I buy something
[18:03:34] <JT-Shop> must have been someone else on the mailing list asking for it
[18:03:41] <gene__> what I want is an absolute lock on z that has to be manuallo unlocked to adjust it
[18:04:57] * JT-Shop goes back to making a worm
[18:07:12] <jdhNC> someone here made the pyvcp button. I made the same mistake twice last night.
[18:10:00] <JT-Shop> pyvcp button for touch off Z?
[18:10:18] <cradek> I've learned to read that window, every time, before hitting enter
[18:10:37] <cradek> it says the axis, "to workpiece"/"to fixture", and radius or diameter if lathe x
[18:10:55] <jdhNC> JT: right.
[18:10:56] <JT-Shop> what I forget to to is change the radio selection then only look at the words and then don't read them
[18:11:03] <JT-Shop> yea, that be mee
[18:11:10] <cradek> any of those things can be a screwup, but it SAYS everything
[18:11:30] <JT-Shop> anyway the touch off button makes it a one mouse click deal for me...
[18:11:38] <jdhNC> oh, you are supposed to read them?
[18:12:05] <cradek> well if you want to be sure you're doing the right thing, it's good to do
[18:12:09] <JT-Shop> only until your fingers deveope their own mind
[18:12:29] <jdhNC> a good part of what I do at work is making it easy for people to do the right thing
[18:12:44] <cradek> sounds hard
[18:12:51] <jdhNC> yep
[18:13:12] <jdhNC> though the hardest part is being forced to attend Human Performance training.
[18:19:58] <JT-Shop> hmm I've got some 18-20 gauge sheet metal somewhere around here for the worm
[18:21:21] <andypugh> cradek: Somebody suggested having to select which axis and what, with it defaulting to "nothing". I have some sympathy with the idea. As a bonus, you could choose to zero all axes in all systems, and the tool table, if you had taken leave of your senses.
[18:21:31] <djdelorie> separate buttons?
[18:21:59] <andypugh> jdhNC: They assess how good your performance at being human is?
[18:22:01] <cradek> adding lots of mouse clicks is an awful solution
[18:22:20] <andypugh> cradek: We aren't all as smart as you :-)
[18:22:30] <cradek> I use the keyboard shortcuts exclusively
[18:22:36] <andypugh> See!
[18:22:42] <cradek> it's a bug in AXIS if you must use the mouse for something
[18:22:53] <jdhNC> andypugh: mostly covering (repeatedly) that people are fallible and if you expect them to do the correct thing every time you are screwed.
[18:23:30] <andypugh> cradek:
http://clock.msurma.net/
[18:23:33] <cradek> to touch off X, hit X, [End], type your value, READ THE SCREEN, hit return
[18:24:05] <cradek> if you want to always select the axis before you touch off, then do that
[18:24:11] <andypugh> My keyboard has no "END". More troublesomely, it has no "Ctrl"
[18:24:21] <cradek> then it's defective
[18:24:21] <jdhNC> you need a new keyboard
[18:24:47] <cradek> andypugh: I just get a weird error
[18:24:55] <andypugh> It's a stainless steel "Kiosk" keyboard, and I think they decided "Ctrl" was dangerous.
[18:24:56] <cradek> "Error #1065"
[18:25:13] <djdelorie> without ctrl, you can't do ctrl-alt-delete
[18:25:22] <andypugh> cradek: With that web site?
[18:26:04] <andypugh> cradek: Maybe
http://esquire.ru/clock ?
[18:27:27] <andypugh> Basically a 7-segment LED clock, done with 6' planks, bolts, ladders and a bunch of people over 24 hours.
[18:28:01] <cradek> ah, I've seen a little clip of it, but I've never seen it tell the right time
[18:28:13] <gene__> I just made a beautiful thread, and found another gotcha in G76
[18:28:30] <andypugh> It might only catch 8 time zones
[18:30:06] <gene__> If you have a following move back to the start position, that move starts and move about .1" before X has backed away from the threads! Moral, in your G76 wrapper, after the G76, do the x move first, then the Z.
[18:33:07] <JT-Shop> it should end at the drive line which should clear the threads
[18:35:32] <tjb1> ethernet wire does not hold up well to soldering
[18:35:52] <andypugh> There are two things about G76 which are a little wierd. There is that initial in-and out, which I think was Cradek's little joke to give us palpitations, and the fact that that return pass has a taper, so that you tend to catch the back of the tool threading tight bores.
[18:35:53] <jdhNC> I soldered some cat5 parts on to an IR sensor last night.
[18:37:58] <JT-Shop> I wondered what that dance move was just before the thread starts
[18:38:22] <tjb1> I'm soldering up a rgb lcd
[18:38:30] <JT-Shop> I never noticed the taper
[18:39:02] <JT-Shop> crap I guess I'll beat the parts out of some 16 ga stock...
[18:41:06] <Tom_itx> what's a good gage to use for an enclosure?
[18:41:10] <andypugh> JT-Shop: The taper is so that the X-infeed at the start of the thread is always the same. Personally I think that is unecessary, as X stops dead anyway prior to threading. (In fact both axes do, waiting for the index).
[18:41:15] <Tom_itx> i may end up making one
[18:41:21] <andypugh> Tom_itx: 1/4"
[18:41:35] <Tom_itx> i'm not planning to shoot at it
[18:41:43] <andypugh> Maybe 1/2" for things more than 12" cube.
[18:42:11] <andypugh> Ah, you should have said, I assumed you were American, so needed to be proof against common firearms.
[18:42:43] <Tom_itx> i figured around 18 or 20 ga maybe
[18:43:08] <Tom_itx> what are pc cases in general?
[18:43:16] <djdelorie> "flimsy"
[18:43:25] <andypugh> You can probably find specs for what the likes of Rittal use at various sizes
[18:45:06] <gene__> Andy, I have noticed an occasional small move inward during the retrace while it is mid run, and the backout at the end of the cut seems to have about a 2 thou overshoot, the handle comes to a stop but turns back inward about 10 degrees
[18:45:44] <skunkKandT> gene__: you are using software stepgen - right?
[18:45:50] <alex_joni> anyone speaking regex ?
[18:46:06] <andypugh> Tom_itx: Not easy to tell directly, but a Rittal 400x400x250 is 13kg
[18:46:21] <gene__> but te backout when G67 is done is not sufficient to clear the threads if you have a combined move back to the origin after the G76, so they must be separated.
[18:46:23] <djdelorie> alex_joni: yes
[18:46:31] <skunkKandT> Try upping you stepgen max accelleration.. It should be about 20% higher - but try maybe 50%
[18:46:38] <gene__> gah G67 s/bv G76
[18:46:50] <skunkKandT> (than your axis accelleration)
[18:47:05] <andypugh> gene__: G76 ought to stop back where it started.
[18:48:30] <skunkKandT> It has been a long time since I have seen it - but I have seen an axis overshot when the stepgen acceleration is set too close to the actual axis max acceleration
[18:49:31] <andypugh> Tom_itx: So, thats 3.6kg.m^2
[18:50:10] <andypugh> 8kg.m^3...
[18:50:35] <andypugh> ~Err, 8000 that is
[18:51:23] <Tom_itx> mine probably doesn't need to be as heavy as a normal industrial enclosure
[18:52:23] <Tom_itx> and i'll have to use something i have available as well
[18:52:46] <Tom_itx> we have a decent aluminum as well as iron scrap yard here
[18:53:11] <Tom_itx> the iron yard doesn't usually have alot of sheet though
[18:54:12] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/patio/patio22.jpg
[18:54:22] <Tom_itx> i drug that nice i beam home under the s10 :D
[18:54:30] <Tom_itx> as well as the posts
[18:54:47] <andypugh> Tom_itx: I am getting 2.2.. thickness for that size of industrial cabinet
[18:54:58] <Tom_itx> mm i presume?
[18:55:14] <Tom_itx> that's rather thick for my needs i think
[18:55:24] <Tom_itx> sounds about right though
[18:55:26] <andypugh> Yes
[18:55:55] <andypugh> What size are you looking at?
[18:56:07] <Tom_itx> box size?
[18:56:10] <andypugh> Yes
[18:56:25] <Tom_itx> i considered an electrical pannel but they're not deep enough
[18:56:44] <Tom_itx> something around 18x24x 6 or 8" deep
[18:57:37] <Tom_itx> the aluminum surplus has some nice stamped boxes but they have no front
[18:57:49] <Tom_itx> i used a smaller one of those on my pendant
[18:57:53] <djdelorie> two electrical panels? ;-)
[18:58:36] <Tom_itx> i'll keep my eye out for something
[18:58:48] <Tom_itx> maybe a mains switch or something may be deeper
[18:59:01] <Tom_itx> or a large starter box
[18:59:32] <djdelorie> sounds about the same size as a 4U enclosure or full tower PC case
[18:59:42] <Tom_itx> close
[19:00:07] <Tom_itx> i'd like a front access hinged door though
[19:00:56] <jdhNC> I was planning on putting my drives/PS/daughterboards in an old PC chassis. By the time I drilled all the cheap rivets out, the case didn't have much left as a case.
[19:01:05] <andypugh> Tom_itx:
http://www.ebay.com/itm/400296596662
[19:02:04] <andypugh> There may be cheaper, but the search was polluted by plccentre and their no-picture, just the part number thing.
[19:03:02] <Tom_itx> yeah something like that would be ok
[19:03:08] <jdhNC> I hate plccenter
[19:03:23] <Tom_itx> i gotta lay stuff out and make sure what size i'm after
[19:05:22] <andypugh> I can't search ebay US by "cheapest". It makes me include postage, which massively distorts things as many items are $20 + $250 shipping
[19:05:45] <Tom_itx> yeah i can do the footwork
[19:06:50] <andypugh> For example this nice 24 x 24 x 8 cabinet is $125 + $686 shipping.
http://www.ebay.com/itm/230781734143
[19:06:53] <jdhNC> I offered them $75 for that case.
[19:07:50] <andypugh> Actually, that last one is nice, even including DIN rail and panduit to make Jymmm happy
[19:08:06] <Tom_itx> we're all about that :D
[19:08:24] <jdhNC> I bought some panduit from McMaster last week. THey shipped one piece in a 5"x8ft tube
[19:08:44] <Tom_itx> extra shipping charge for the long package?
[19:09:10] <jdhNC> yeah, they shipped it alone. $19 for the panduit, $19 for shipping. Should have gotten two
[19:09:13] <andypugh> Panduit is horrible, learn harness lacing :-)
[19:10:13] <Tom_itx> i posted a couple links on that the other day
[19:11:06] <jdhNC> http://www.ebay.com/itm/400265126593
[19:11:13] <jdhNC> kind of shallow though... nice price.
[19:11:58] <Tom_itx> 6" might be ok for my needs
[19:12:10] <Tom_itx> i need to lay stuff out and see though
[19:13:19] <Tom_itx> would you stick the heatsink fins out the side or leave them in and put a fan on them:
http://tom-itx.dyndns.org:81/~webpage/cnc/psu/heatsink2.jpg
[19:14:08] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/cnc/psu/driver_wiring.jpg
[19:14:46] <andypugh> Tom_itx: I might just leave them inside with no fan as a Plan A. The box surface area is orders of magnitude greater than that of the components you are trying to cool.
[19:15:00] <Tom_itx> yeah
[19:15:22] <andypugh> Yay! cable lacing:
http://www.frobenius.com/bletchley/bl032.jpg Bonus points for naming the device.
[19:15:50] <Tom_itx> holy crap
[19:15:58] <Tom_itx> that's gotta be phone equipment
[19:16:11] <andypugh> No.
[19:16:21] <andypugh> There is a clue in the URL
[19:16:34] <jdhNC> it's a Bletchley!
[19:16:45] <andypugh> It's _at_ Bletchley
[19:16:56] <Tom_itx> what the heck are all the little gears for?
[19:17:21] <jdhNC> mechanical computer?
[19:17:48] <andypugh> http://www.frobenius.com/bletchley/bl035.jpg
[19:18:16] <andypugh> http://www.frobenius.com/bletchley/bl038.jpg
[19:18:38] <jdhNC> a turing bombe rebuild!
[19:18:39] <Tom_itx> http://www.dairiki.org/hammond/cable-lacing-howto/
[19:18:42] <andypugh> http://www.frobenius.com/bletchley/bl039.jpg
[19:19:10] <Jymmm> somebody call me?
[19:19:26] <Tom_itx> nope
[19:19:29] <jdhNC> people were calling you names but I told them to stop.
[19:19:31] <andypugh> jdhNC:
http://en.wikipedia.org/wiki/Bombe (Or how the Poles won the war)
[19:19:43] <Tom_itx> http://www.apartmenttherapy.com/try-cable-lacing-to-get-loose-97907
[19:19:45] <Jymmm> jdhNC: no fun in that
[19:19:54] <Jymmm> andypugh: panduit?
[19:20:09] <Jymmm> never heard of it,
[19:20:21] * Tom_itx gives Jymmm some long shoe laces and sends him off to learn about cable lacing
[19:23:30] <Jymmm> When I worked at ____, one of the amp cabinets was just amazing. you could VISUALLY trace every wire in ever harness from point to point. The harness were self-supported. The water cooling hoses didn't use one clamp and never leaked. It was THE MOST BEAUTIFUL wiring job I have ever seen. Far more a work of art than anything else.
[19:25:12] <Jymmm> The amp was in 3 19" rack type cabinets, so you can imagine how much cabling there was.
[19:26:47] <Jymmm> Does anyone know about themo sealing plastics by chance? Specifically PP
[19:27:44] <r00t4rd3d> http://www.robotshop.com/diamondback-arduino-wifi-microcontroller-2.html
[19:28:45] <andypugh> Jymmm: No, I only know anythng about doing it deliberately.
[19:29:00] <Jymmm> andypugh: Heh
[19:35:13] <gene__> Andy, thanks for the help, its working pretty good now. But the temps are dropping like a rock, so I'm outta here
[19:36:46] <djdelorie> maybe you should feed them more... ;-)
[19:37:05] <jdhNC> nah, that's why you hire contractors
[19:43:13] <Jymmm> maybe not feed them , but at least give them some brown water
[19:48:27] <alex_joni> Jymmm: found a couple more infected files
[19:50:33] <tjb1> Got the 16x2 lcd working, gonna use it to display tip voltage :)
[19:53:49] <Jymmm> alex_joni: Any obvious pattern to them?
[19:54:35] <Jymmm> alex_joni: I'd RO the files you've cleaned (at least for now), till you are sure you've cleaned everything.
[19:55:56] <alex_joni> they are all RO
[19:56:02] <alex_joni> now
[19:56:12] <Jymmm> alex_joni: cool
[20:03:41] <ReadError_> alex_joni
[20:03:45] <ReadError_> run clamav
[20:03:54] <ReadError_> its pretty good about find bad php/shell scripts
[20:04:12] <ReadError_> its typically what i use at work
[20:06:35] <tjb1> Should this be purchased with a new tig welder -
http://www.harrisweldingsupplies.com/lincoln-parts-kit-kp508.aspx
[20:07:11] <alex_joni> ReadError_: not installed on that box
[20:08:41] <ReadError_> you just have a shell?
[20:08:58] <ReadError_> you can mount it as an SCP share on a local box
[20:09:02] <ReadError_> and run it that way ;)
[20:09:07] <ReadError_> somewhat ghetto but should work
[20:09:18] <ReadError_> SFTP*
[20:11:44] <alex_joni> yeah, I'll give it a go tomorrow
[20:12:49] <Jymmm> alex_joni: You'll need to RTFM ClamAV btw.
[20:13:08] <Jymmm> alex_joni: It's not "straight forward" intuitive as one may think.
[20:13:39] <Jymmm> alex_joni: LOTS of options when running/updating it
[20:13:40] <alex_joni> Jymmm: surely not for 4am ;)
[20:14:00] <Jymmm> alex_joni: No, not for 4am, more like 1pm
[20:14:09] <alex_joni> siesta :D
[20:14:25] <Jymmm> Buenos Nachos and Tacos too!
[20:14:48] <alex_joni> yeah, that
[20:15:49] <Jymmm> G'Night MrAlex
[20:19:41] <ReadError_> Jymmm
[20:19:47] <ReadError_> i run freshclam
[20:19:48] <ReadError_> then this:
[20:20:28] <ReadError_> ./usr/bin/clamscan -r -i --log=/var/log/clamav/scan.log /path/to/scan
[20:23:11] <Jymmm> ReadError_: Right, but standard nomenclature would be clamav - update, not a seperate update utility =)
[20:24:31] <ReadError_> thats how they do i think
[20:24:40] <ReadError_> well atleast in the EPEL repo releases
[20:34:59] <toast2> turns out, kevlar thread is really slippery and doesn't seem to hold knots well
[20:35:37] <Jymmm> toast2: what are you knotting it for?
[20:36:05] <Jymmm> usually used in weaved products that I've seen, then sewn in layers of soemthign else.
[20:36:18] <djdelorie> when I last used kevlar cord, for model rocketry, I made sure all the knots were inside the rocket and embedded in epoxy.
[20:36:38] <Jymmm> djdelorie: for the chute?
[20:36:41] <djdelorie> yup
[20:37:08] <Jymmm> that reminds me, I need to pick up some phenolic
[20:37:17] <ReadError_> hmmm
[20:37:22] <ReadError_> so visualmill is really nice
[20:37:24] <djdelorie> two strands, folded in the middle to make two loops, then the four ends were braided together to form a cord. The end of the cord was deep inside the rocket, leaving only a loop on the outside
[20:37:26] <Jymmm> Gawd that's gonna be expensive
[20:37:27] <ReadError_> i just need to learn it better
[20:37:30] <ReadError_> still my 1st day with it
[20:37:59] <djdelorie> not for the chute cords themselves, for the tether that the shock cord is connected to
[20:38:11] <Jymmm> between each stage?
[20:38:25] <djdelorie> kevlar - shock cord - chute cords - chute
[20:38:30] <Jymmm> k
[20:38:41] <djdelorie> with the nosecone or upper section in there somewhere ;-)
[20:39:03] <Jymmm> you hope =)
[20:39:42] <Jymmm> 3 2 1... clic...SON OF A BITCH! I knew I forgot something *SWISH*
[20:39:57] <Jymmm> too late now gomer!!!
[20:40:41] <djdelorie> http://www.delorie.com/photos/20041002-cmr814/IMG_3980.JPG
[20:41:12] <Jymmm> Yep, there be the *SWISH*
[20:41:36] <djdelorie> that's my biggest, 9 feet tall, three E9 engines
[20:41:36] <ReadError_> anyone here us a ps3 controller as a pendant?
[20:41:41] <ReadError_> i know alex4nder does
[20:41:45] <ReadError_> but anyone else?
[20:43:30] <Thetawaves> djdelorie, i'm surprised you don't make your own motors.
[20:43:56] <djdelorie> there's an excellent write-up in the rec.model.rockets FAQ about why that's a bad idea, written by a friend of mine.
[20:44:05] <Thetawaves> rocket candy is pretty benign, though not so powerful
[20:44:37] <Jymmm> I know where you can get some Potasium Nitrate if so inclined
[20:44:48] <djdelorie> http://www.ninfinger.org/rockets/rmrfaq.5.html#q18
[20:46:23] <Jymmm> djdelorie: Wuss!
[20:46:44] <djdelorie> a wuss with three 160 volt servo motors in his cnc machine :-)
[20:46:55] <Thetawaves> seems like that faq amounts to 'waaaa it's too hard'
[20:47:11] <djdelorie> read the last paragraph
[20:49:47] <Thetawaves> most of it's lost on me, i'll be a hundred feet away from it
[20:50:06] <alex4nder> hey
[20:50:42] <ReadError_> hey alex4nder
[20:50:47] <Thetawaves> but i have worked quite a bit with rocket candy and it is ... weak stuff
[20:50:50] <Thetawaves> difficult to light
[20:50:56] <ReadError_> you use the ps3 controller with that qtsixa ?
[20:51:07] <alex4nder> no
[20:51:09] <ReadError_> it sees it, gets the #s and all
[20:51:14] <ReadError_> but i think its for BT only
[20:51:16] <alex4nder> I just use hal_input and X11's joystick
[20:51:27] <ReadError_> got a copy of you config?
[20:51:39] <alex4nder> back in the lab, but it's easy to setup
[20:51:43] <alex4nder> just man joystick
[20:51:46] <alex4nder> and look at the hal_input examples
[20:51:52] <alex4nder> also I'd suggest installing jstest-gtk
[20:53:00] * Thetawaves is one of those guys more interested in how the motor works than how the rocket flies
[20:53:56] <alex4nder> ReadError_: the only thing that's weird is: you need the hid-sony quirk to turn the PS3 controller 'on'
[20:57:10] <ReadError_> finally got some solid toolpaths alex4nder
[20:57:15] <ReadError_> visualmill ;)
[20:57:24] <alex4nder> what's their removal strategy?
[20:57:35] <ReadError_> very flexible
[20:57:37] * djdelorie enjoys the technical challenge of the build
[20:57:43] <alex4nder> ReadError_: but what is it?
[20:57:51] <ReadError_> i have no idea ;)
[20:58:01] <ReadError_> its got a bunch of different options
[20:58:06] <ReadError_> pocketing, profiling, drilling
[20:58:28] <alex4nder> yah, this is where the difference comes into play
[20:58:44] <alex4nder> let me know how it works out
[20:58:51] <alex4nder> if you remember, make some vidoes
[20:59:51] <ReadError_> of the cuts?
[21:00:10] <alex4nder> yup
[21:00:11] <ReadError_> the only thing i cant figure out, is when i home, then touchoff
[21:00:15] <ReadError_> my z is super high
[21:00:21] <ReadError_> but it will still run
[21:00:28] <ReadError_> im using g54
[21:00:42] <alex4nder> where does your CAM reference its zero? top of the material, or bottom?
[21:00:53] <ReadError_> i have it at the bottom
[21:01:05] <alex4nder> and you're touching off Z there?
[21:01:10] <ReadError_> if i have it at the top, it thinks its negative
[21:01:15] <ReadError_> and says nugusta
[21:01:32] <ReadError_> well, z from the top of the material
[21:01:39] <ReadError_> but
[21:01:39] <alex4nder> so what happens if you move Z to where it thinks 0 is, and then touch off with an offset of 0?
[21:01:42] <ReadError_> i have to account for thickness
[21:02:05] <alex4nder> your CAM software is accounting for thickness though, right?
[21:02:05] <ReadError_> well when i say its high, its like out of bounds high
[21:02:11] <ReadError_> yea
[21:02:12] <ReadError_> i set all that right
[21:02:19] <ReadError_> i just cant find where the offset is
[21:03:28] <alex4nder> what I do is I home the mill, fixture the material, and then move the spindle to where I know exactly where it is in relationship to what I'm working on, and then touch off.
[21:03:32] <ReadError_> hmm actually
[21:03:39] <ReadError_> on this machine, its fine
[21:03:47] <ReadError_> i was running in vmware esxi for testing earlier
[21:03:51] <ReadError_> maybe i need to resync my configs
[21:04:17] <ReadError_> i also got all the vectric stuff
[21:04:29] <ReadError_> cut2d, cut3d, vcarve, the photo one
[21:08:03] <Solaris> Beddy byes for snoozey pies!!!
[22:15:07] <toastydeath> do any of you guys do hand woodworking
[22:17:39] <djdelorie> define "hand"
[22:17:45] <ReadError_> is that some kind of innuendo ?
[22:17:46] <ReadError_> hehe
[22:17:53] <ReadError_> working wood with your hand...
[22:17:59] <toastydeath> jack planes, chisels, back saws
[22:18:00] * ReadError_ has been on the internet too long
[22:18:07] <toastydeath> hand powered tools
[22:18:10] <elmo40> by hand do you mean clamp it to a vice and press Start? ;)
[22:18:26] <djdelorie> I do a little, but usually as part of a larger power-tool project
[22:19:41] <toastydeath> i am looking at buying nothing but hand tools because it's all the room i've got
[22:19:56] <ReadError_> you need a bunk bed
[22:20:00] <jdhNC> all you need is a good knife.
[22:20:01] <ReadError_> more room for activities
[22:20:39] <ScribbleJ> toastydeath, how do you plug yourself into linuxcnc? I'm confused.
[22:20:56] <toastydeath> ?
[22:21:03] <Tom_itx> heh
[22:21:30] <elmo40> ScribbleJ: no plug required. Bluetooth all the way! ;)
[22:21:56] <Tom_itx> good set of chisels and a mallet
[22:22:19] <djdelorie> and a ryoba
[22:22:20] <ScribbleJ> elmo40, if you have a bluetooth you might want to see a dentist.
[22:22:32] <ReadError_> so yall
[22:22:40] <ReadError_> i ordered a d525
[22:22:44] <ReadError_> should be here tomorrow
[22:22:56] <ReadError_> if it sucks, ill turn it into a media pc or something and get the one alex4nder has
[22:23:29] <elmo40> ReadError_: what does alex4nder have?
[22:23:30] <jdhNC> I have an intel d525
[22:23:41] <ReadError_> 4700 something i think
[22:23:45] <ReadError_> i dont have it on this pc here
[22:23:52] <ReadError_> jdhNC, does it suck?
[22:24:02] <jdhNC> nope, great. small, cheap, low latency
[22:24:09] <Tom_itx> it's a good board for linuxcnc
[22:24:10] <ReadError_> what do you get ?
[22:24:13] <Tom_itx> i have one
[22:24:21] <ReadError_> turns out, perfect timing
[22:24:30] <ReadError_> just put 16gb ddr3 in my macbook
[22:24:38] <ReadError_> so i have some dimms for it ;)
[22:24:55] <jdhNC> I only put 2gb in mine, seems fine.
[22:25:08] <Tom_itx> i filled mine, whatever it would hold
[22:25:11] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/intel_atom/atom_index.php
[22:25:29] <Tom_itx> newegg is your friend
[22:25:32] <ReadError_> i think 4gb is the limit
[22:25:38] <ReadError_> i got an amazon gift card from work
[22:25:42] <ReadError_> so i used it
[22:25:49] <ReadError_> plus, amazon prime i got overnight for 3.99
[22:25:52] <jdhNC> I bought mine with an amazon gift card from work
[22:26:25] <ReadError_> ;o
[22:26:27] <ReadError_> lol
[22:26:35] <ReadError_> what kind of latency yall getting?
[22:26:43] <ReadError_> alex4nder was getting 7k with his
[22:26:45] <ReadError_> or something super low
[22:26:53] <Tom_itx> sounds about right
[22:27:12] <jdhNC> I can get mine up higher with IO
[22:27:27] <jdhNC> but, I don't and I'm using a mesa so I don't really care.
[22:27:40] <Tom_itx> same here
[22:27:43] <jdhNC> what size wire do people use for steppers
[22:28:09] <Tom_itx> i'm not sure what i had
[22:28:35] <ReadError_> oh you get about 7k also?
[22:28:49] <Tom_itx> gawd i gotta reset my router...
[22:28:51] <jdhNC> 5k just sitting there
[23:55:35] <joe9> is an intel atom cpu the recommended cpu for linux cnc?
[23:55:51] <joe9> or, is there something better/cheaper?
[23:56:01] <Thetawaves> cheaper than an atom?
[23:56:06] <joe9> i have an amd one, but that has latency issues once in a while.
[23:56:18] <Thetawaves> yeah thats bad
[23:56:26] <joe9> and a celeron cpu that gives overruns within a few mins of starting the latency test.
[23:56:56] <joe9> i am not sure if I can. but, it would be cool if the intel atom cpu is compatible with the celeron.
[23:57:01] <joe9> will have to check.
[23:57:16] <joe9> Thetawaves: do you recommend the intel atom?
[23:57:26] <joe9> i have never used it. but, I have heard it mentioned here before.
[23:57:32] <Thetawaves> i have one, works fantastic
[23:57:45] <joe9> ok, cool. will get that. thanks for the advice.
[23:57:58] <Thetawaves> maximum latency is about 7us
[23:58:57] <Thetawaves> i have Atom 330 2x
