Back
[00:01:53] <CaptHindsight> tjtr33: ever work with
http://www.smaltec.com/ Micro-EDM?
[00:08:37] <tjtr33> CaptHindsight, no, interesting Lisle. I've used Agies micro and seen the Sarex and Panasonics. funny a new guy just miles away
[00:11:20] <tjtr33> an funnier, that is a really precise spindle, damn simple but very true.
[00:11:38] <tjtr33> gnite for now :)
[01:36:51] <gabewillen> memleak in the new kernel
[02:05:21] <DJ9DJ> moin
[05:06:44] <RyanS> Got myself an el cheapo 4x6 bandsaw ... hope it's a time saver
[05:49:33] <Jymmm> JT-Shop: cradek: pcw_home: Dollar Tree has these car to USB (1A) adapters. I cracked one open an it has one of these
http://www.ti.com/product/mc34063a Input 3-40VDCm Output 1.25-40VDC @ 1.5A. they might be perfect to provide 5VDC from various other sources and the PCB is tiny.
[05:49:59] <Jymmm> s/DCm/DC,/
[06:11:04] <herron> probably not isolated
[06:12:59] <archivist_herron> fleabay 170734419936 for similar
[08:32:16] <cradek> skunkworks:
https://www.youtube.com/watch?v=mRru1bKxd2c
[08:32:28] <cradek> at 0:11 you can see what the old control looked like
[08:32:50] <cradek> interesting style of program listing
[08:41:04] <skunkworks> heh - I think I have watched almost all the emco compact5 lathe videos on youtube.
[08:41:19] <skunkworks> we have one 'cnc' version like that - but the thing doesn't boot.
[08:42:56] <skunkworks> cradek, I almost have a hal file hacked togather.. I used lut and comp for setting the switch velocity. (I also had to use abs... )
[08:43:23] <skunkworks> the pc software actually looks pretty cool - it shows the cutter and stock for the compact 5
[08:45:00] <skunkworks> https://www.youtube.com/watch?v=wQoWnB3oByo
[08:47:25] <PetefromTn> Morning Folks
[08:49:13] <skunkworks> cradek, I am going to just use a fixed offset between the half and full stepgens first. If I need a little lag/lead between them - (a step or so) I can change it so negative velocites get one offset and positive get another..
[08:49:39] <skunkworks> it is really seeming like it will work. ;)
[08:54:20] <skunkworks> cradek, heh
https://www.youtube.com/watch?v=Ztuv71S5eVs
[09:33:49] <fbx90> skunkworks: nice sweater tie combo that guy has
[09:37:50] <cradek> skunkworks: looks like quite a piece of software
[11:47:39] <cbjamo> I need to pre-process my gcode, and want to use a python script to do so, can I assign a filter to run a specific script on the input file?
[11:48:48] <cradek> http://www.linuxcnc.org/docs/html/gui/axis.html#_program_filters
[11:52:44] <cbjamo> So I could have for example PROGRAM_EXTENSION = .gcode then gcode = gcodepreprocessor
[11:53:40] <cbjamo> Then I would need a program called gcodepreprocessor in the path? and how would it pass the input file to the program?
[11:58:35] <jthornton> look at the example python script
[12:01:24] <CaptHindsight> interesting topic on the user list about the size of G-code files for GGG's since they have a 0.3mm tool/nozzle building parts 3 orders of magnitude larger than the dia
[12:02:44] <`Nerobro__> CaptHindsight, huh?
[12:03:18] <CaptHindsight> this gets further complicated if printheads or a micro-valve array is used since they might only have 25um nozzles
[12:03:30] <cbjamo> CaptHindsight: that is definately a problem, I have a small, if complex part, that is ~300,000 lines
[12:04:23] <cbjamo> I understand that I could run the script directly and have that select the gcode I need to procsess, but I would rather select the file from the axis gui and have the script run without user interaction.
[12:04:28] <`Nerobro__> Aaah, that makes sense
[12:04:39] <`Nerobro__> Perhaps gcode isn't the way to program 3d printed objects?
[12:05:03] <CaptHindsight> the issues is do you control the actual deposition with G-code or just the motion with G-code?
[12:05:25] <cbjamo> CaptHindsight: I am controlling the deposition with the gcode.
[12:05:33] <CaptHindsight> maybe G-code just for the motion control
[12:05:50] <cbjamo> but then how do you sync the motion with the extrusion?
[12:06:13] <CaptHindsight> and syncing the deposition from a frame buffer
[12:06:29] <CaptHindsight> like an inkjet
[12:07:13] <CaptHindsight> the g-code shrinks
[12:07:47] <CaptHindsight> it's simple if you print 2.5D like a reprap since you can have a bitmap for each layer
[12:08:15] <CaptHindsight> it gets more complicated if you print in 3D
[12:08:18] <cbjamo> indeed, although an object like this one will still be huge,
http://www.thingiverse.com/thing:35246
[12:09:05] <`Nerobro__> CaptHindsight, that seems more sane. T
[12:09:22] <`Nerobro__> But I don't know how you'd get the plastic grain right
[12:09:36] <`Nerobro__> but a bitmap seems like the best plan for something like laser sintering.
[12:09:47] <`Nerobro__> or glueing powder.
[12:10:03] <CaptHindsight> DLP, inkjet and e-beam
[12:11:53] <CaptHindsight> the problem with having only one small nozzle (besides having only one small nozzle) is that it has to travel over the entire part, same for only one laser
[12:13:02] <CaptHindsight> so the G-code for motion is large and if the part has lots of on/off for deposition it's really large
[12:15:53] <`Nerobro__> other than potential throughput issues... what's the 'big" deal if a g-code file is big?
[12:16:49] <CaptHindsight> what about a few years from now?
[12:17:22] <CaptHindsight> if you can print sub-micron and the part is 500mm^3
[12:18:09] <CaptHindsight> maybe a vector format or lossless compression scheme for the deposition
[12:18:21] <archivist> the main issue id the entire file gets checked when loading
[12:18:26] <archivist> is
[12:18:48] <CaptHindsight> the worst part will be parts like sponges or foam, lost of on/off
[12:19:15] <archivist> or some form of colour/grey scale too
[12:19:19] <CaptHindsight> but even RLE for GGG's would compress most part files
[12:21:14] <CaptHindsight> SVG's for DLP work except we had some issues using ARM's since the libs weren't very accurate
[12:23:49] <IchGuckLive> hi all B)
[12:25:16] <CaptHindsight> Was ist los?
[12:28:41] <DJ9DJ> whuzzup? ;)
[20:55:08] <Maggio> Hello everyone, and thank you for being there. I started to read the user manual, which is impressionnat! Congratulations. Before I plunged more details. I want to know if I could get a first result with a PC equipped with a joystick and a DB25 port, and one or two small stepper motor connected to a wire / resistance?
[20:58:21] <Tom_itx> you would need hbridge drivers for the motors
[20:58:36] <Tom_itx> not sure about the joystick interface but something could be done
[20:59:02] <Tom_itx> the parport can put out step and direction signals for the steppers
[21:11:36] <Maggio> Tom_itx: Thank you for you reply. I am very new. I checked on the H-Bridge drivers and it does not seem difficult. But it does not correspond to the idea that I was doing ... I will continue my reading because i never use stepper motor
[23:12:27] <chopper79> PCW around?
[23:13:32] <pcw_home> yeah
[23:13:44] <chopper79> Have a couple questions if you have some time?
[23:14:13] <pcw_home> sure
[23:15:36] <chopper79> hm2_[HOSTMOT2](BOARD).0.gpio.002 this would be an output pin according to dmesg. If this is true then why would I recieve an error when launching linuxcnc stating that pin hm2_[HOSTMOT2](BOARD).0.gpio.002 does not exist?
[23:15:46] <chopper79> 5i25 board
[23:16:29] <chopper79> net xstep => hm2_[HOSTMOT2](BOARD).0.gpio.002 guess I should have put this part also
[23:17:10] <pcw_home> It will not exist if its being used as a step/dir/pwm etc output
[23:17:58] <chopper79> ok it is an output just an invisible one
[23:18:24] <chopper79> how does it not exist but exist?
[23:19:59] <pcw_home> Because its owned by the channel 0 stepgen (so not available as GPIO)
[23:20:53] <chopper79> ok, this is why I can setp for this output but can not link it to anything else.
[23:21:18] <pcw_home> so if you set num_stepgens to 0 it will be available (but you will have to hardware stepgens)
[23:21:58] <chopper79> Well as you can tell im still trying to work on that squaring issue I have been working on. Still can not find a way to square the gantry
[23:22:38] <pcw_home> JA3 is probably the way (and working for other gantry users)
[23:22:43] <chopper79> ja3 and gantrykins have not worked out well for me
[23:23:25] <chopper79> This would be fine if it was something that was serviceable over the phone if something happened in the field
[23:23:40] <pcw_home> gantrykins is not what you want
[23:24:10] <chopper79> Gantrykins is for sure not what I want
[23:24:39] <chopper79> I have had my fill of it and do not care to mess with it anymore
[23:25:16] <pcw_home> maybe best to ask micges about ja3 (which looks like it may be default for the upcoming 2.6 release)
[23:26:20] <chopper79> Sound good, I know the last time me and him had a conversation I was able to get it fired up but as I said it would be a horrible thing to service over the phone with anyone if something happened to the computer.
[23:26:26] <chopper79> When is 2.6 coming?
[23:27:08] <pcw_home> probably best asked on the devel channel
[23:27:24] <chopper79> got ya
[23:28:09] <chopper79> Would a tandem output and squaring be something that you could program on the 5i25?
[23:28:36] <chopper79> I thing the tandem motors could be but the squaring maybe not
[23:28:42] <chopper79> *thnk
[23:28:45] <chopper79> think
[23:30:49] <pcw_home> Anythings possible but this is really a LinuxCNC gantry homing issue so thats where its best fixed
[23:31:40] <pcw_home> I do know that micges had production gantry machines running JA3
[23:32:03] <pcw_home> so its certainly do-able
[23:32:45] <chopper79> I understand
[23:33:01] <chopper79> Well sir, thanks for your time and have a good night
[23:34:12] <pcw_home> Its should be easier when 2.6 is released
[23:37:19] <chopper79> hope so