Back
[06:08:16] <jthornton> M1
http://www.usmicroscrew.com/wp-content/uploads/pdf-download-icon.png
[06:09:22] <jthornton> http://jimorrisco.com/index/thread-info
[08:17:30] <archivist> http://www.nczonline.net/blog/2015/05/14/the-bunny-theory-of-code/
[08:50:35] <jthornton> expanding html docs
http://gnipsel.com/linuxcnc/html/
[08:55:53] <archivist> hmm title-content possible mismatch
http://gnipsel.com/linuxcnc/html/common/Integrator_Concepts.html
[08:56:48] <archivist> no mention of other realtime stuff and other concepts a user needs
[08:57:26] <archivist> was just picking a random page from the indec
[08:57:29] <archivist> index
[08:59:16] <jthornton> that is a copy of our current 2.7 docs and yes perhaps mislabeled which is my next step after we make a decision about the pdf docs
[08:59:19] <archivist> eg integrator either uses a wizard to connect stuff (stepconf) or edits hal
[09:00:03] <jthornton> in my thinking an integrator is the person that builds the machine
[09:01:17] <archivist> me too
[09:04:02] <jthornton> the main thing I'm showing here is how much neater and compact the html docs can be and everything shows up on one page with a "normal" monitor to start
[09:06:14] <archivist> cant help looking deeper when I look at something :)
[09:06:21] <REEEN> Testing
[09:06:28] <archivist> phail
[09:07:55] <jthornton> no problem I appreciate the feedback
[10:49:39] <REEEN> Hello, some skilled users online ?
[10:52:12] <micges> what's the problem?
[10:53:12] <REEEN> I tried to tune a costumers mill but there is a problem with the tuning, there is always a offset present
[10:54:19] <REEEN> X axes has no offset but y and z are always exact 0.004 mm and 0.008mm away from position inpositiv direction
[10:55:18] <micges> why are you tunning them?
[10:55:57] <micges> I mean new machine, retrofit, or simply retune?
[10:56:08] <REEEN> I did the pid tuning i did a retrofit on this machine
[10:56:22] <REEEN> Yes retrofit
[10:56:47] <REEEN> I did sone retrofits before,bjt neber had such a problem
[10:58:40] <micges> did you replace servo drivers?
[10:59:38] <REEEN> No I used the old Bosch Servo driver DC powered and the old Siemens DC servo motors they are in very good condition
[11:01:30] <REEEN> For the x axes my tuning is perefect, but y has an offset of 0.004 and z by 0.008 so if I say Mdi G0 Y0 display says Y = 0.004 and PID Error can also be seen in hal meter window
[11:05:42] <cradek> is that inches or mm? how many encoder counts are in .004 units?
[11:07:10] <REEEN> It's mm it is very tiny but on my other retrofits I didn't even had a offset
[11:07:28] <cradek> .004 mm?
[11:07:40] <REEEN> 0.004 is 4 encoder counts I use linear glass scales
[11:07:50] <cradek> what are your tuning parameters currently?
[11:07:52] <REEEN> Yes 0.004 mm
[11:08:04] <REEEN> And in z 0.008 mm
[11:08:32] <cradek> are these velocity mode amps?
[11:09:08] <REEEN> Yes PWM analog control
[11:09:26] <REEEN> Amplifiers are dc
[11:11:32] <cradek> what are your tuning parameters currently?
[11:11:43] <cradek> are these velocity mode amps?
[11:11:49] <REEEN> So you mean PID values ?
[11:11:59] <cradek> yes pid
[11:12:03] <REEEN> Yes
[11:12:53] <cradek> ok glass scales go to linuxcnc; do the amps have tach or encoder feedback?
[11:13:02] <REEEN> P =22 I = 0.00258 D =0.003 and FF1= 0.1 or so
[11:13:35] <REEEN> The amps have tach feedback and scales go to linuxcnc directly you are right
[11:13:39] <cradek> ok
[11:13:53] <cradek> how is ferror when moving?
[11:14:48] <REEEN> Max ferror during Rapid movement is less than 0.006 added to the permanent offset
[11:14:58] <REEEN> It is very good
[11:15:21] <cradek> yes 6 encoder counts at rapid is very good
[11:15:39] <cradek> I recommend setting I=0 and fixing the offset with the knob on the amp
[11:15:53] <REEEN> Yes of course there is a higher peak at the start of the movement
[11:15:55] <jepler> just a fyi -- I will have very little IRC contact for the next ~1month though I'm likely to continue idling in the channel.
[11:16:39] <REEEN> I think you are right cradek I thought the same
[11:17:22] <REEEN> I recognized that the offset decreases after some time but very very slow
[11:17:40] <cradek> that is your very small I building up
[11:17:57] <cradek> try setting I=0 and tune the offset on the amp
[11:18:27] <REEEN> Yes I will try that out :-)
[11:18:28] <cradek> once you have done that you can set I=100 or I=1000 safely, but you probably don't need it
[11:18:41] <cradek> do NOT use I when you are adjusting the offset because it will confuse you
[11:18:57] <REEEN> I have a lot of other machines that don't have any I
[11:19:16] <REEEN> Only p and FF1
[11:19:17] <cradek> in velocity mode, I=0 makes perfect sense
[11:19:24] <cradek> yes and maybe D
[11:19:36] <REEEN> Yes and D of course :)
[11:19:42] <cradek> right, it should mostly rely on FF1
[11:19:54] <REEEN> Yes it does
[11:20:21] <REEEN> You are a skilled tuber I have one more question for you if you have a moment
[11:21:12] <cradek> haha skilled tuber. hot potato.
[11:22:07] <REEEN> The machine is moving very hard from position to position when using a handwheel with 0.1mm per click how can I make this smoother ?
[11:22:40] <REEEN> Turning down accel is very bad for positioning it gets tooo slow
[11:23:22] <jepler> there is a component 'ilowpass' for this purpose
[11:23:23] <cradek> you can fix that with ilowpass
[11:23:40] <cradek> http://linuxcnc.org/docs/html/man/man9/ilowpass.9.html
[11:23:49] <jepler> ah cradek was a bit faster with the doc link
[11:23:59] <REEEN> Ah okay so this is what this can be used for :)
[11:24:51] <REEEN> So using ilowpass will only smooth movement when using handwheel are for all jog moves ?
[11:25:14] <cradek> it smooths only the handwheel
[11:25:37] <REEEN> Okay but thats enough
[11:25:42] <REEEN> Thank you
[11:25:58] <REEEN> Thanks for your help buddys :)
[11:26:11] <cradek> there's an example in ./common/shuttlexpress.hal
[11:26:24] <cradek> actually there are several ilowpass wheel setups in the sample configs
[11:26:32] <cradek> welcome!
[11:27:02] <REEEN> Okay I will find out how to use it Iam pretty skilled in hal things
[11:27:13] <cradek> excellent
[11:27:24] <REEEN> Have a good day !
[11:27:24] <cradek> it's so much easier to help people when they've done their homework :-)
[11:27:26] <cradek> you too.
[11:27:35] <REEEN> Yes of course :)
[11:28:02] <micges> REEEN: if you're connecting handwheel directly to axis.n you can try also set axis.%d.jog-vel-mode to 1
[11:28:50] <cradek> that will not make it less responsive
[11:29:38] <REEEN> Yes but this will make my machine go on positioning also when I stopped turning my handwheel I think this is to dangerous
[11:30:20] <micges> I see
[11:47:02] <archivist> cradek, I cant help thinking your screw is some 00 coarse relation of Sellers standard
[11:54:53] <kwallace> I'd like to write a Python script that modifies an .ini file, such as find [TAG], then keyword = ?? then write a new value to the keyword. Is there an easy way to do this or do I just need to use standard Python file and string functions?
[12:01:46] <cradek> from linuxcnc import ini
[12:01:57] <cradek> ^ I see the linuxcnc module has ini stuff, but I don't know if you can write/rewrite with it
[12:03:22] <cradek> there is also ConfigParser
[12:03:49] <cradek> you absolutely shouldn't write this yourself from scratch.
[12:04:08] <kwallace> ConfigParser is in C?
[12:05:48] <kwallace> Touchy has this:
[12:05:51] <kwallace> 256 # check the ini file if UNITS are set to mm"
[12:05:51] <kwallace> 257 # first check the global settings
[12:05:51] <kwallace> 258 units=self.ini.find("TRAJ","LINEAR_UNITS")
[12:15:07] <kwallace> hmm,
http://linuxcnc.org/docs/html/common/python-interface.html#_reading_ini_file_values_a_id_sec_python_reading_ini_values_a
[12:18:09] <jepler> kwallace: ConfigParser is a standard Python module
[12:18:26] <jepler> but no guarantees that our inifiles are 100% compatible with ConfigParser's idea of inifiles
[12:18:39] <jepler> we also have several homebrewed variants of inifile reader/writers in tcl and they are all buggy
[12:18:59] <jepler> dgarr fixed a couple of bugs in the one in calibration within the last few months
[12:19:35] <kwallace> Cool, thank you and cradek. I'll look into it.
[12:38:24] <pcw_home> $39.00 ebay DC7800SFF works pretty well with hm2_eth at 3 KHz (wont quite do 4 KHz with normal desktop activities), 2 weeks 24/7 no issues
[12:38:25] <pcw_home> RX bytes:516632415322 (516.6 GB) TX bytes:902056655900 (902.0 GB)
[12:38:27] <pcw_home>
[13:06:54] <skunkworks> neat
[13:08:46] <JT-Shop> kwallace, I've done some ini file creating and reading in python
[13:11:42] <JT-Shop> kwallace, look in the first part of
http://gnipsel.com/files/mill-g-code/
[15:04:58] <mozmck> pcw_home: is the 7i92 with 2 IDC headers close to ready?
[15:45:09] <mozmck> cmorley: can I change keybingdings easily in a gscreen custom skin?
[15:45:22] <cmorley> yes
[15:45:58] <mozmck> Ok, I want to change the jog keys to call functions in my custom screen.
[15:46:08] <mozmck> how would I do that?
[15:47:32] <cmorley> this is the principle:
http://www.linuxcnc.org/docs/devel/html/gui/gscreen.html#_use_the_handler_file_to_add_features
[15:47:51] <cmorley> I will look for a more detailed example....
[15:48:29] <cmorley> It's been awhile since I looked at it..
[15:48:54] <mozmck> thanks. I'm finding more in the code for gscreen now - so I can probably sort it out.
[15:49:18] <mozmck> I want to make key jogging use a percentage of the max velocity for each axis.
[15:49:50] <mozmck> Looks like it should be easy, but I have to implement my own key jog functions.
[15:49:51] <cmorley> keybindings.py will show you the built in keybuindings
[15:49:59] <mozmck> I found that.
[15:50:22] <mozmck> Made me learn a little more python to try and understand it :)
[15:52:39] <cmorley> :)
[15:52:45] <cmorley> still looking..
[16:01:00] <mozmck> I think that example has what I need. I really want to override the gscreen key jog functions instead of changing keybindings.
[16:03:39] <cmorley> self.gscreen.keylookup.add_call('XPOS','new_up_key_pressed_function')
[16:04:48] <cmorley> should be the command you need to have the up button rediredted to the function 'new_up_key_pressed' that would need to be added to the handler file
[16:05:07] <mozmck> From the link you sent, it looks like all I need is a on_keycall_XPOS(self,state,SHIFT,CNTRL,ALT) function in my code to override the one in gscreen?
[16:06:16] <cmorley> need the add_call command too
[16:06:22] <mozmck> ok
[16:07:00] <mozmck> so will that override the previous add_call that bound it to on_keycall_XPOS()?
[16:07:12] <cmorley> sorry my memory of it off hand is fuzzy :)
[16:07:20] <cmorley> yes it should
[16:07:27] <mozmck> ok, I'll give that a try.
[16:14:56] <cmorley> I tested here and yes I can override the default action for a jog key
[16:27:44] <cmorley> i'd like to see your gui sometime, if you had screenshots.
[16:34:54] <cmorley> just to confuse you a bit - you were right the add_call is not necessary - but won't hurt - it does allow you to rename your new function to what you want.
[17:48:34] <mozmck> cmorley: thanks for the help.
[17:48:52] <mozmck> Right now my gui is still a mess, but I'll have screenshots at least before too long.
[17:51:15] <andypugh> Who is admin for the sourceforge site?
[17:51:35] <andypugh> https://sourceforge.net/p/emc/feature-requests/new/ has a 404 on the “Before submitting a Feature Request, you should read this webpage.” link
[18:28:12] <cradek> andypugh: looks like I can... I removed that. If you think there's an appropriate link, I can put it back.
[18:28:13] <andypugh> I don’t know where it was meant to link to
[20:04:56] <cradek> andypugh: the thing about setting new reports to priority 1 was particularly stupid, because we can choose the default and it's 5
[20:05:06] <cradek> andypugh: ... so I just deleted both sentences
[22:54:20] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/threadtiming 918c618 06linuxcnc 10src/hal/hal_lib.c 10src/hal/hal_priv.h 10src/hal/utils/halcmd_commands.c 10tests/save.0/expected hal_lib: new pin,param for thread time,tmax * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=918c618
