Back
[04:45:56] <REEEN> Hello, would be nice if some developers are only, I would like to discuss some missing features of linuxcnc
[04:46:05] <REEEN> Safety based features
[04:47:33] <REEEN> Iam doing retrofits on a wide range of machines and setting up a good hal for safety is pretty complicated due to following errors
[04:53:24] <micges> what specific features are missing?
[04:57:09] <REEEN> There is no possibility to lock a single axes in motion I have several cases where this is very important
[04:58:01] <REEEN> The only way to block the motion in hal leads to a lot of problems with following errors and so on
[05:00:10] <REEEN> Just that you understand me right, I want to disable an axes and if I try to move it there should be a message like axes is locked an no occuring following error
[05:00:56] <micges> I understand
[05:03:08] <REEEN> Iam also using some amps that provide a feature called sls (safe limited speed) if I request this feature my axes will not be allowed to move faster than a safe speed, this is for machines with cabins, if you open the door axes will only be allowed to move with a safe speed, same problem here, no way to tell linuxcnc that axes are not allowed to move fastee
[05:04:44] <micges> you can use motion.adaptive-feed for that along with cabin door sensors
[05:04:58] <REEEN> This is a very very important thing to make cnc machines safe. Machines are not even allowed to move faster with open doors, if someone gets injured the machine builder has to pay for that
[05:05:13] <REEEN> Really ?
[05:05:51] <micges> adaptive-feed scales whove velocity down from 100% to 0% and it's realtime
[05:06:14] <micges> whole
[05:06:33] <REEEN> Okay that solution is good but not perfect at all
[05:07:11] <REEEN> Does adaptiv feed also scale jogs ?
[05:08:04] <micges> I know SLS from drives should be safer but I would make retrofit setup with only one source of command velocity
[05:09:12] <REEEN> Yes I do it with the drives but the problem was that I can't tell linuxcnc : you are not allowed to move faster !
[05:10:04] <micges> there is also max velocity input in halui
[05:10:22] <REEEN> The problem is that software based "sls" is not conform
[05:11:34] <REEEN> Yes thank you
[05:12:44] <REEEN> I wrote a conponent to open the doors and place the machine in a safe state, it does the following:
[05:16:39] <REEEN> It pauses the programm if any then motion.feed and motion.speed inhibit are activated, another special component measures the speeds of my axes and spindle(it is not possible to use encoder velocity because of pid peaks) and if all axes are stopped I assert the break in z and disable all amps and then I power off the drives, the power of and the door switch unlocking is done with a safe plc. That is only conform way to safely ope
[05:18:22] <micges> I would also use motion inhibit input from drives
[05:18:32] <REEEN> I read a lot about safety with cnc machines so I really know what is right and what is wrong for CE or UL
[05:18:45] <REEEN> Yes this happens by hardware
[05:19:24] <REEEN> The drives are powered off so they are really safe
[05:20:08] <REEEN> Thanks to feed and speed inhibit the user can try to move the machine with no occuring following errors
[05:20:45] <REEEN> But there are cases I have to block motion of only one or 2 axes not all at once
[05:23:54] <micges> there is enable/disable command in emcmot, probably making hal pins for that would help
[05:24:13] <micges> you could disable specific axis
[05:24:19] <REEEN> Yes that would be awesome :)
[05:25:51] <micges> fix is easy but tests to see if nothing was broken would be hard
[05:26:44] <REEEN> Yes this probably harder
[05:28:25] <REEEN> Maybe we can make a feature request for that ?
[05:28:36] <micges> yeah
[06:50:24] <ikcalB> CaptHindsight: you already left a note for ikcalB (reading the logs - have to set up my irc proxy finally...)
[06:52:03] <ikcalB> jepler: my nick is registered with nickserv. is that enough, for '/msg memoserv' to work?
[07:02:25] <ikcalB> CaptHindsight said something about the new RTAI implementation - it wouldn't use legacy i-Pipes anymore. Are you going to change linuxcnc / keep maintaining NTULINUX/RTAI?
[07:02:57] <ikcalB> and what about other RT systems anyways
[07:55:20] <jepler> we welcome implementations of RTAPI for other RTOSes but they so seldom show up in mergeable form.
[08:19:26] <jepler> personally I think it would be interesting to have a port to RTAI LXRT userspace and have started toying with it, but I can't promise any results.
[09:29:35] <jepler> spindle "homing" and spindle in position mode -- now that master branch is foremost in our minds, how should we approach it?
[09:32:56] <CaptHindsight> ikcalB: memleaks RTAI already supports 3.16 and 3.18 kernels with linuxcnc
[09:35:59] <CaptHindsight> ikcalB: changing linuxcnc is up to the devs here, you'll have to ask them about that
[09:39:48] <jepler> I'd be happy to review any proposed patches to support new RTOSes or newer versions of RTAI
[09:57:32] <ikcalB> CaptHindsight: Thank you very much Sir! your input saved me loads of trouble!
[10:38:58] <skunkworks> wow - that velocity command doesn't make sense..
[10:43:41] <skunkworks> oh - is he going in and out - that would be. (sort of making a spooler)
[10:49:02] <seb_kuzminsky> skunkworks: your microphone is on
[10:49:27] <skunkworks> heh
[11:24:48] <KGB-linuxcnc> 052.6-for-jessie e981f08 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e981f08
[11:28:34] <KGB-linuxcnc> 052.7-for-jessie 9811b63 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9811b63
[11:38:56] <KGB-linuxcnc> 05jepler/hm2-eth 93fc025 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=93fc025
[11:38:56] <KGB-linuxcnc> 05jepler/hm2-eth-multiple 2a83a1b 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2a83a1b
[11:38:56] <KGB-linuxcnc> 05jepler/hm2-eth-packetreduction da07d64 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=da07d64
[11:38:57] <KGB-linuxcnc> 05jepler/hm2-eth-simplify 66f0669 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=66f0669
[11:38:59] <KGB-linuxcnc> 05jepler/hm2-eth-startup c02b348 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c02b348
[11:39:03] <KGB-linuxcnc> 05jepler/hm2-eth-v2 d95d654 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d95d654
[11:39:07] <KGB-linuxcnc> 05jepler/hm2-eth-v3 c72834b 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c72834b
[11:39:10] <KGB-linuxcnc> 05jepler/hm2-resolver-no-stepgen eb5f2c7 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=eb5f2c7
[11:40:31] <skunkworks> clean up, clean up, every one - everywhere
[11:41:43] <KGB-linuxcnc> 05jepler/2.6/halcompile-extra-link-args 3991f9c 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3991f9c
[11:41:43] <KGB-linuxcnc> 05jepler/arm-atomic a694daf 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a694daf
[11:41:43] <KGB-linuxcnc> 05jepler/elbpcom 45cf42e 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=45cf42e
[11:41:45] <KGB-linuxcnc> 05jepler/libreadline-gplv2-v2.6 79c1921 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=79c1921
[11:58:19] <KGB-linuxcnc> 03Jeff Epler 05master 6534839 06linuxcnc 03scripts/checkglade checkglade: a new script to check glade files * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6534839
[11:58:19] <KGB-linuxcnc> 03Jeff Epler 05master c813031 06linuxcnc 10src/emc/usr_intf/stepconf/Submakefile stepconf build: refactor to repeat less * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c813031
[11:58:19] <KGB-linuxcnc> 03Jeff Epler 05master e2bace3 06linuxcnc 10src/emc/usr_intf/stepconf/Submakefile stepconf: non-fatal check for duplicate IDs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e2bace3
[12:01:53] <KGB-linuxcnc> 03Jeff Epler 05master 7417d4a 06linuxcnc 10src/Makefile 10src/configure.in 10src/rtapi/uspace_rtapi_app.cc Merge remote-tracking branch 'origin/jepler/aarch64' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7417d4a
[12:01:59] <KGB-linuxcnc> 05jepler/2.6/checkglade 62684b9 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=62684b9
[12:01:59] <KGB-linuxcnc> 05jepler/aarch64 f1c84aa 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f1c84aa
[12:02:33] <KGB-linuxcnc> 05jepler/2.6/proposed bd2577c 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bd2577c
[12:03:32] <KGB-linuxcnc> 05jepler/hardcode-cpuid-for-ch 2edbb6b 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2edbb6b
[12:03:32] <KGB-linuxcnc> 05jepler/hm2-encoder-dpll 357ce42 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=357ce42
[12:03:32] <KGB-linuxcnc> 05jepler/rtos-master-v0-linux_rt 383baf7 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=383baf7
[12:04:24] <KGB-linuxcnc> 05jepler/uspace-pci-sysfs-enabled 52bfb20 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=52bfb20
[12:05:13] <KGB-linuxcnc> 05jepler/check-tclsh-load 5d8d360 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5d8d360
[12:10:50] <jepler> REEEN, it sounds like your goal is to sell a LinuxCNC retrofit service. In that case, you should do more than request that features be added by volunteer programmers: you should find software developers to pay to make the improvements you select in a way that we can integrate those changes into the LinuxCNC that is shipped by linuxcnc.org.
[12:12:02] <jepler> we are happy to give feedback to devlopers you hire and pay about proposed patches so that they can refine them and make them suitable for inclusion in linuxcnc.org's software.
[12:22:54] <KGB-linuxcnc> 03Jeff Epler 05jepler/hal-streams 1613fb5 06linuxcnc 10(10 files in 3 dirs) hal: factor out streamer/sampler * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1613fb5
[12:22:54] <KGB-linuxcnc> 03Jeff Epler 05jepler/hal-streams 8406bbb 06linuxcnc 03docs/man/man3/rtapi_stream.3rtapi hal: document the new C API * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8406bbb
[12:23:15] <jepler> well that's just about it for my tidying up
[12:23:33] <jepler> I hope to do something interesting with hal-streams soon
[12:58:43] <skunkworks> zlog
[13:38:52] <seb_kuzminsky> jepler: awesome
[14:44:51] <seb_kuzminsky> tus .
[14:55:39] <cradek> seb_kuzminsky: thanks for massaging the bug tracker
[15:02:21] <skunkworks> does the current livecd have the mesa firmware?
[15:02:27] <skunkworks> (or how would I know?)
[15:05:24] <cradek> because the linuxcnc package "recommends" hostmot2-firmware-all, they are on the cd
[15:05:45] <skunkworks> awesome
[15:06:02] <cradek> how would you know: you could do dpkg -l 'hostmot2*' or look in synaptic
[15:06:38] <cradek> now there are firmwares that are not pacakged at all, like the ones for the 5i25 etc
[15:07:08] <cradek> these packaged ones are the ones the hostmot2 driver uses to program the cards
[15:07:51] <cradek> http://paste.ubuntu.com/12330599/
[15:13:09] <skunkworks> right - I need the classic ones. :) 5i20...
[15:24:26] <cradek> man who hasn't thought about cutting a knurl with a threading tool
[15:24:41] <cradek> ... and then got a knurling tool after trying it
[16:03:13] <seb_kuzminsky> always a nice surprise when i look at the bug tracker and realized we've actually fixed things people reported
[16:03:34] <cradek> sometimes it's not even an accident
[16:04:54] <seb_kuzminsky> and sometimes (like now) i try to verify that something's gotten fixed, and get sidetracked because i discover another problem while testing the first problem
[16:06:16] <cradek> it's turtles all the way down
[16:07:37] <seb_kuzminsky> https://lh3.googleusercontent.com/-EsGSqkQW9j0/UKU4M0LgGRI/AAAAAAAAJfU/gsHGSCuyXoo/w1125-h844-no/12%2B-%2B1
[16:12:06] <skunkworks__> http://i.imgur.com/nggMDfE.png
[16:15:15] <cradek> skunky: you guys updated the K&T?
[16:27:26] <seb_kuzminsky> guess that makes you the last of the old guard, cradek :-P
[16:30:10] <cradek> ?
[16:30:56] <cradek> oh you mean I should update jr?
[16:30:59] <cradek> yeah I wonder what it's running
[16:32:13] <seb_kuzminsky> soemthing that aint broke, i guess
[16:32:22] <cradek> surely
[16:32:26] <cradek> it works every time
[16:32:48] <cradek> I could find another CF card and try a debian+2.7 install
[16:32:56] <cradek> I think it's new enough to boot from USB so it'd be easy
[16:33:18] <seb_kuzminsky> it runs off CF? that's surprising, i never had good lifetime from those
[16:33:35] <cradek> my desktop at home boots from CF too
[16:33:39] <cradek> I've also used them in laptops
[16:34:02] <cradek> I bought a SSD for the desktop, it lasted a few months and bricked, so I put the CF card back
[16:34:09] * cradek shrugs
[16:36:48] <skunkworks> zlog
[16:37:53] <skunkworks> yes - the harddrive was going. Dad got an ssd to put in it.
[16:38:04] <skunkworks> only issue was getting the arduino running.
[16:43:12] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 b679ccc 06linuxcnc 03tests/loadrt.2/expected 03tests/loadrt.2/loadrt.hal 03tests/loadrt.2/test.sh tests: add another loadrt test * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b679ccc
[16:43:12] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 b4eb195 06linuxcnc 10src/rtapi/sim_rtapi_app.cc sim_rtapi_app: clean up on failed "realtime" module load * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b4eb195
[16:48:48] <skunkworks> I think it is the first time it has run the installed version!
[16:49:06] <cradek> heh
[16:49:13] <cradek> yay for packages
[17:16:03] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 84b3873 06linuxcnc 03src/rtapi/uspace_rtapi_app.cc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=84b3873
[17:16:03] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 5467503 06linuxcnc 10tests/loadrt.2/expected tests: add new streamer pins to loadrt test * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5467503
[17:17:45] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master aa37cb0 06linuxcnc 10debian/changelog 10src/rtapi/uspace_rtapi_app.cc Merge remote-tracking branch 'origin/2.7' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=aa37cb0
[18:20:34] <andypugh> PCW: It looks like the new lathe will be using resolvers. I have a 6i25 and a 7i80DB. Will either work with a 7i49? You did mention a DB25 version of the 7i49 a little while back?
[18:21:04] <andypugh> Otherwise I guess that 5i24 + 7i49 is the cost-effective setup?
[18:43:27] <skunkworks> zlog
[22:01:08] <skunksleep> Great quote... It (mach4) will support whatever the motion device can support...
[22:33:55] <seb_kuzminsky> i wonder how linuxcnc would do as a "motion device" for mach4
[22:36:22] <skunksleep> Hmmm...,
