Back
[06:42:59] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/moveoff e1bce10 06linuxcnc 10configs/sim/axis/moveoff/hookup_moveoff.tcl hookup_moveoff: support 9 axes (personality=) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e1bce10
[09:35:11] <pcw_home> peak hm2-eth read times (above minimum) reduced from 114 usec to 27 usec by setting Ethernet IRQ priority!
[09:36:03] <cradek> ooh
[09:36:53] <skunkworks> How doe one set irq priority?
[09:36:58] <pcw_home> after ~24 hours of youtube videos (which normally trigger the occasional peaks)
[09:37:06] <pcw_home> chrt
[10:00:58] <skunkworks> wow gene
[10:01:53] <cradek> sigh
[10:02:59] <skunkworks> I have not been following much - but is he trying to do cutter comp in gcode?
[10:03:27] <cradek> I don't know
[10:03:35] <cradek> I never seem to understand his troubles
[10:04:06] <cradek> I think lately it comes down to how he just doesn't get how to write an IJ arc
[10:04:16] <skunkworks> cradek, did you see
http://www.electronicsam.com/images/Mazak/comu.JPG
[10:04:18] <Connor> skunkworks: reminder: db 25 + db 9 pc back plate :)
[10:04:19] <cradek> a lot of people don't get it
[10:04:25] <skunkworks> reason for broken brushes...
[10:04:32] <cradek> a therefore a lot of people think IJ arcs are buggy or something
[10:04:37] <skunkworks> Connor, let me look
[10:05:20] <Connor> skunkworks: Wow.. that's pretty gnarly.
[10:05:21] <cradek> why is it bulgey? is one of them loose?
[10:06:14] <skunkworks> yes - a section delaminated
[10:07:42] <pcw_home> really looks like oil or coolant contamination
[10:14:00] <skunkworks> Connor, how many do you need? (I have only found 1 so far...)
[10:16:42] <pcw_home> skunkworks: sudo chrt -p 99 (your RT Ethernet IRQ PID)
[10:30:19] <Connor> skunkworks: Just 1 with both 25 and 9.
[10:35:10] <skunkworks> Connor,
http://electronicsam.com/images/KandT/testing/1208141011.jpg
[10:35:19] <Connor> Yup.
[10:36:01] <Connor> I had access to 100's of those things back when I worked in a local PC Store and repair shop.
[10:36:05] <Connor> they're long gone now. :(
[10:37:20] <skunkworks> I could throw it in a usps envelope
[10:41:14] <Connor> If you wouldn't mind. I'll cover postage.
[10:41:22] <Connor> I.E. paypal you
[10:43:30] <skunkworks> eh - don't worry about it.
[10:45:37] <CaptHindsight> they are still on ebay
[10:46:03] <Connor> CaptHindsight: Yea. But, they charge a arm and a leg for shipping..
[10:46:25] <CaptHindsight> I can't belive I had to buy new ones after the 100's I had tossed into boxes years ago
[10:46:37] <Connor> I know? Right?
[10:48:10] <cradek> be aware those are both serial connectors
[10:58:02] <pcw_home> also be aware that there are 2 different serial port pinouts used on motherboards
[10:58:54] <pcw_home> one is IDC connector --> flat cable compatible the other is not
[10:58:56] <Connor> pcw_home: Don't need the serial connectors. just the bracket.
[10:59:06] <archivist> does anyone keep to proper serial port specs!
[10:59:50] <Connor> the db 25 is going to be replaced with full 25 pin db 25, the db 9 is being replaced with hd db 15 (vga)
[11:00:32] <pcw_home> Ahh no serial
[11:01:47] <pcw_home> archivist: this is the 10 pin motherboard header pinout I was talking about some are 1-1 other are IDC compatible
[11:36:43] <archivist> I am used to the "fun" getting serial working with sensible handshake
[11:49:08] <seb_kuzminsky> did you guys see dgarr's latest moveoff video? It now does waypoint backtracking (optionally).
[11:49:25] <seb_kuzminsky> https://www.youtube.com/watch?v=KY6hx7WBkO8
[11:54:26] <skunkworks> it isn't good enough
[11:54:30] <skunkworks> ;)
[11:54:40] <skunkworks> Pretty darn neat
[12:29:52] <Connor> how is moveoff different than jog while pause ?
[12:30:52] <Connor> or is that what we're calling it for now ?
[12:31:58] <skunkworks> different implimtation
[12:32:18] <Connor> I can't keep it all straight anymore.
[13:10:45] <seb_kuzminsky> Connor: moveoff is a hal-only trick to let you move the tool while you have a program paused
[13:11:19] <seb_kuzminsky> it provides the same functionality as the jog-while-paused patch that was floating around, but without polluting our core code with interim solutions
[18:14:32] <mozmck> Are userspace hal components run as separate processes?
[18:15:00] <cradek> yes
[18:16:54] <mozmck> Ok, so the main HAL process calls functions to update pins every thread period, but the component can have separate loops and timers doing whatever.
[18:17:27] <cradek> yes userspace components are free-running and there is no telling when they'll read and write their pins
[18:19:08] <mozmck> ok, I think I understand.
[18:20:16] <mozmck> So if I wanted several components to talk to each other then I would need some sort of IPC
[18:20:40] <mozmck> I think I'll be better off making one big component.
[20:35:41] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/moveoff 2156370 06linuxcnc 10configs/sim/axis/moveoff/hookup_moveoff.tcl 10configs/sim/axis/moveoff/moveoff.hal 10src/hal/components/moveoff.comp moveoff: not_done_yet note for addf work needed * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2156370
[20:35:41] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/moveoff 57a18e1 06linuxcnc 10(13 files in 2 dirs) moveoff.comp: Separate functions for read, write * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=57a18e1
[21:30:07] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 8172042 06linuxcnc 10debian/control.in 10scripts/latency-histogram latency-histogram: clean up on ^C * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8172042
