#emc | Logs for 2010-11-30

Back
[00:00:25] <emcrules> not really after i fixed it i felt like my 7 year old
[00:03:51] <theorb> theorb is now known as theorbtwo
[00:10:04] <JT-Hardinge> I was playing 7 years old during Thanksgiving with my grand kids
[00:10:11] <JT-Hardinge> it is fun
[00:36:13] <Valen> what is a decent manufacturer for spindles in the 3KW or so class
[00:49:08] <emcrules> GC coloumbo
[00:49:19] <emcrules> or PDS
[00:49:34] <emcrules> for spindles
[00:59:48] <andypugh> There was a chap on cnczone who bought a cheap Chinese one and seemed very pleased indeed with the quality. I am guessing it wasn't one of the millions on eBay, but who can say?
[01:05:24] <andypugh> This is the thread. http://www.cnczone.com/forums/benchtop_machines/77851-bt30_spindle_-_tool_changing_switch.html
[01:05:57] <andypugh> I don't think I would install one of those if I was being paid for the job though.
[01:18:00] <emcrules> How can i simulate my z axis for the time being so i can run my mill in x and y only?
[01:23:18] <andypugh> what do you need to simulate?
[01:25:28] <emcrules> i dont have a z axis yet but would like to run programs
[01:26:03] <emcrules> but i do have my z axis setup in the ini and hal files
[01:26:37] <emcrules> needless to say when i run a program that call for a z movement i get following errors
[01:27:13] <andypugh> Servo or stepper?
[01:28:32] <andypugh> I assume servo (with a stepper you can just not enable to drives).
[01:30:11] <emcrules> yes servo
[01:30:13] <andypugh> You could use the sim_encoder hal function to give a convincing feedback, but it is probably easier to just loop-back axis.2.motor-pos-cmd to axis.2.motor-pos-fb in the HAL file (commenting out the real links for the time being, and assuming that Z is axis.2
[01:30:17] <andypugh> http://www.linuxcnc.org/docview/html//man/man9/sim_encoder.9.html
[01:30:57] <emcrules> i will try that thanks.
[01:31:55] <andypugh> basically comment out the line that has axis.2.motor-pos-fb on it, and copy the motor-pos-fb to the end of the line with motor-pos-cmd in it
[01:33:22] <andypugh> You will also need to comment out the lines in the INI file which define the homing speed, direction and sequence. (maybe just change the LABEL names, as I don't think the concept of commenting out really applies to the iNI file)
[01:34:01] <andypugh> Night all.
[01:41:21] <juri_> well, i now have a table the size of the side of a full tower case, that moves in one axis, at 160 stops, on 16TPI threaded rod.
[01:42:15] <juri_> its got an arm mounted above it, for my Y and X axises to be mounted to.
[01:42:55] <juri_> er. Y and Z.
[03:29:24] <emcrules> Has anyone used the facing py routine found on the wikki
[04:07:05] <atmega> noy mr.
[04:07:09] <atmega> err... not me.
[05:31:19] <elmo401> elmo401 is now known as elmo40
[05:45:11] <Howard> Hi All,
[05:47:38] <Howard> Hi All, I ma trying to write a M1XX file to read a pin and exit a while loop when the state change from 1 to 0 ?
[05:48:26] <Howard> Hi psha, how are you doing ?
[05:49:25] <Howard> I am trying to write a M1XX file to read a pin and exit a while loop when the state change from 1 to 0 ?
[05:51:00] <psha> and what's wrong?
[05:51:01] <psha> hi
[05:51:01] <the_wench> hello psha, you have a question?
[05:51:23] <psha> the_wench: yea, main question is what i'm doing here so early...
[05:52:35] <Howard> How about some thing like this . a = net pin10 parport.0.pin-10-in
[05:53:03] <psha> while [ "x`halcmd getp pin`" != "x0" ]; do sleep 1; done
[05:53:23] <psha> something like this wil lhelp you :)
[05:53:48] <Howard> AH ! thanks psha :0)
[05:54:18] <Howard> What time is it over there ?
[05:54:51] <psha> not earyly :) 08.50
[05:55:57] <Howard> What is the temp like ?
[05:56:31] <psha> -15 :)
[05:57:24] <psha> nice weather for bicycle :) not too cold and not much ice on the roads :)
[05:57:28] <Howard> Oh shit. We are suffering 25c here and it is 5.00 pm
[05:58:55] <psha> my wife would be happy if we have 25 here :) but for me it's good when its cold:)
[06:01:07] <Howard> I was work in Germany in the early 70 and it was -35c that kills if you are too long out in it.
[06:01:56] <Howard> I guess you would be normal in winter in Moscow ?
[06:03:31] <psha> heh... good weather (<-20) are rare in moscow lately...
[06:03:56] <psha> it's too large city with too many 'heaters' like cars, poorly insulated buildings etc...
[06:04:40] <psha> so there is at least +5 diff with temp in any place 50-100 km from moscow
[06:06:09] <Howard> This is where I live on the coast. http://www.portmacquariedirectory.com.au/
[06:06:20] <psha> -35 is excelent weather for kids since some schools are closed and they may play in the street for a whole day :))
[06:08:12] <Howard> They must be tough kids at that temp or really good gear on them.
[06:09:48] <psha> they are active :) when you are active there is no difference (to some extent) what's temp
[06:11:49] <Howard> Do you live in the centre of Moscow ?
[06:12:48] <psha> was living in the center, now somewhere in 2/3 distance to the border
[06:13:25] <Howard> Which border is that ?
[06:14:14] <Howard> I am 400 Km North of Sydney and about 1 Km from the coast.
[06:14:31] <psha> of moscow border ;)
[06:14:39] <Howard> OK
[06:15:19] <Howard> Are you emploted in the CNC game ?
[06:15:24] <psha> closest coast is baltic sea (~720km)
[06:15:39] <psha> no, HPC
[06:16:07] <Howard> HPC ?
[06:16:09] <psha> CNC contract is my side work since main job gives not enought money to live :(
[06:16:24] <psha> high performance computing, distributed systems
[06:16:45] <psha> primary scientific tasks
[06:17:36] <Howard> I've been a programmer for 25 tears mainly C and stand alone system for small machines.
[06:19:02] <Howard> Years not tears :0)
[06:19:14] <psha> heh, i'm working only for ~8 years...
[06:19:26] <psha> first commercial prog was somewhere in 2002
[06:19:50] <Howard> I am nearly 70 ! :0)
[06:20:32] <psha> :)
[06:21:29] <Howard> I was designing computers in 1970 for a German Company and it ran up until 2000 all in TTL
[06:22:21] <psha> nice age for computer...
[06:23:04] <psha> they were more reliable in that time
[06:23:15] <Howard> They were not interested in pulling it out because it kept working OK
[06:23:47] <psha> 'don't touch what is working'
[06:23:49] <psha> :)
[06:24:34] <Howard> If it ain't broke don't fix it ! :0)
[06:25:04] <Howard> I am Howard and you are ?
[06:25:09] <psha> Pavel
[06:25:32] <Howard> OK thanks.
[06:26:06] <Howard> We were were trained on valve technology back in the 60's
[06:27:17] <psha> i was studying math so low-level techs are not common to me :(
[06:27:43] <psha> only recently i've started to study something like PIC's and simple schematics
[06:27:57] <Howard> My son has a Masters degree in computing science
[06:29:16] <psha> it seem that i'm too lazy to do my thesis :)
[06:29:37] <Howard> I have a friend that is a whiz kid on PIC'S and think nothing of writing 1000 line programs in a couple of days.
[06:30:57] <Howard> Are you at work now ?
[06:31:29] <psha> now, at home... i'm sitting with kid in the first middle of the day while my wife is at work
[06:31:46] <psha> i'm programmer and there is no difference where i'm working - at home or at work
[06:31:54] <Howard> How many kids do you have ?
[06:32:04] <Howard> or just the one ?
[06:32:04] <psha> and she is working with cells
[06:32:21] <Howard> cells ?
[06:32:25] <psha> just one till now...
[06:33:08] <psha> cells. she's working in some cancer research
[06:33:09] <Howard> Boy or a Girl ?
[06:33:56] <psha> girl, 2 yrs
[06:34:42] <Howard> I have some very interesting research for her if she is interested and it is giving excellent results.
[06:35:59] <Howard> My son who 30 30 has just had a large brain tumor removed about 75 mm diameter 5 weeks ago and is doing very well.
[06:37:44] <psha> she is working on methods of using own blood cells to carry cancer markers. but i'm not confident in what's she is doing since my education is highly technical :)
[06:38:10] <Howard> There are many people that have survived 10 years and more with this type of tumor "Glioblastoma Multiform" and normall they only last 1 year.
[06:39:34] <Howard> I'll give you my email address if you wish and I can send the data over to you is you wish ?
[06:40:47] <psha> i'll ask her but she's work is mostily in theoretical stage -- considering different methods of binding cancer markers to lymphocytes and not with immuno-related areas
[06:43:53] <Howard> OK but there is about 30 years of research in this area and we know some of the leading people in the area.
[06:44:23] <Howard> here is my email bruhoward@gmail.com
[06:45:13] <psha> thanks, i'll ask her today and report to you tomorrow
[06:46:08] <psha> truly speaking her current work is same as CNC is for me - side projects... here main one was in fundomental science but here lab was closed this year...
[06:46:23] <psha> s/here/her/g
[06:46:40] <Howard> that's great as we are helping a number of people with serious cancer problems.
[06:48:33] <psha> it's good :) it seem that cancer is very widespread nowdays
[06:50:04] <Howard> Yes it is and there are some really simple solution to handle it that have been discovered and I'll send you the link tomorrow.
[06:50:22] <Howard> Do you have Skype ?
[06:51:23] <psha> no, but have gmail
[06:52:12] <psha> lately it's on par with skype with functionality...
[06:53:36] <Howard> I have a camera on Skype and I haven,t used Gmail in this mode as yet.
[06:56:47] <psha> i've registered skype acc some years ago but have not used it... and also have neither video cam nor mic on main computer :)
[06:57:07] <psha> only recently i've acquired one for work on integrating EMC with cameras
[06:57:29] <Howard> OK so we are in chat mode only :0)
[06:58:14] <Howard> What is your band width like over there ?
[06:58:56] <psha> depends on time of day :)
[06:59:17] <psha> in evenings it as low as 4mbit
[06:59:38] <psha> now it's like 30-40mbits
[07:00:55] <psha> heh, tests tell me that it's symmetric 5mbit now...
[07:01:09] <psha> it seem that provider broke something recently :)
[07:02:26] <Howard> That's good as most of the time 30 mbits also
[07:03:35] <Howard> I'll have to disappear for about 30 mins while I get something to eat as it's 6.00 pm here
[07:04:53] <Howard> We are a very large country and the remote areas are usually a bit on the slow side.
[07:06:05] <Howard> Do you know how to setup emc in 6 or 8 axis ?
[07:06:35] <psha> add new [AXIS_X] sections to ini and increase number of axis?
[07:06:45] <psha> and bind relative controls to real pins?
[07:07:12] <Howard> Real pins
[07:07:48] <Howard> I got you now
[07:15:28] <Howard> psha I'll be back ni 20 mins :0)
[07:17:08] <psha> :)
[08:10:19] <Howard> psha are you still around ?
[08:10:25] <psha> yea
[08:11:01] <Howard> You daughter is 2 years old ?
[08:11:26] <Howard> I have a Grand son that is also 2.
[08:19:24] <psha> yea
[08:19:33] <psha> 2 + epsilon
[08:22:57] <Bonny> hello...
[08:24:19] <Bonny> psha do you have time to solve yesterday problems or you have some other important job to do.?
[08:24:37] <Bonny> (wan't to be polite)
[08:24:41] <Howard> I am getting a new machine up and running in the next week or two and I have been working on it for 2 years.
[08:25:10] <Bonny> for wood or metal?
[08:25:18] <Howard> Metal
[08:25:39] <Howard> Not a mill or Lathe
[08:25:52] <Bonny> what then cam be?
[08:25:58] <Bonny> can
[08:26:26] <psha> Bonny: external slider for circle radius?
[08:26:50] <Bonny> psha yes.. and DRO as overlay.
[08:27:09] <psha> for dro - make symlink $EMC2_HOME/include -> emc2
[08:27:11] <Bonny> did you read forum for Andy's idea?
[08:27:40] <Bonny> you talking spanish to me. How to do that?
[08:27:50] <psha> yes, i'll catch him in the evening and we'll talk about it some more
[08:28:00] <psha> ln -s $EMC2_HOME/include emc2
[08:28:06] <psha> in emc-camunits dir
[08:28:11] <psha> then make
[08:28:21] <psha> i'll be back shortly
[08:30:21] <Bonn1> ok
[08:31:24] <Bonn1> http://pastebin.com/EdzwshGG
[08:37:05] <psha> /usr/include/emc2/rtapi.h:752:26: error: linux/ioport.h: No such file or directory
[08:37:24] <psha> it's from emc2-dev from 2.3?
[08:40:31] <Bonny> 2.5.0~pre
[08:41:39] <psha> you have installed package from buildbot?
[08:42:01] <psha> i'm talkin about file /usr/include/emc2/rtapi.h
[08:42:05] <Bonny> What this? I get package with git and compile.
[08:42:18] <psha> so you have emc2-dev package installed or not?
[08:42:26] <psha> check with dpkg -l emc2-dev
[08:42:51] <Bonn1> not found
[08:43:22] <psha> dpkg -S /usr/include/emc2/rtapi.h
[08:44:01] <Bonn1> slavko@emc2:~/emc-camunits$ dpkg -S /usr/include/emc2/rtapi.h
[08:44:01] <Bonn1> dpkg: /usr/include/emc2/rtapi.h not found.
[08:44:31] <psha> but you have one...
[08:45:21] <Bonn1> I have emc 2.4. from live cd installed there and 2.5.0~pre as run in place. But I use just 2.5 version.
[08:45:48] <Bonn1> (I mean ubuntu and emc2.4 are from live CD)
[08:46:01] <Bonn1> I didn't recompile rtapi kernel.
[08:46:10] <psha> it's not rtapi kernel
[08:46:15] <psha> it's header file from somewhere :)
[08:46:42] <Bonn1> seems that I didn't have needed it (yet(
[08:47:57] <Bonn1> I have option to instal emc2-dev
[08:48:13] <Bonn1> it's listed in package manager but not installed.
[08:48:55] <psha> wait a bit, i'll test and fix it
[08:49:04] <Bonn1> ok
[08:50:06] <psha> git pull
[08:51:17] <Bonn1> slavko@emc2:~/emc-camunits$ git pull
[08:51:18] <Bonn1> remote: Counting objects: 5, done.
[08:51:18] <Bonn1> remote: Compressing objects: 100% (3/3), done.
[08:51:18] <Bonn1> remote: Total 3 (delta 2), reused 0 (delta 0)
[08:51:18] <Bonn1> Unpacking objects: 100% (3/3), done.
[08:51:18] <Bonn1> From git://psha.org.ru/psha/emc-camunits
[08:51:18] <Bonn1> 81c339c..74b2c34 master -> origin/master
[08:51:19] <Bonn1> Updating 81c339c..74b2c34
[08:51:19] <Bonn1> Makefile: needs update
[08:51:20] <Bonn1> fatal: Entry 'Makefile' not uptodate. Cannot merge.
[08:51:20] <Bonn1> slavko@emc2:~/emc-camunits$
[08:52:42] <Bonn1> I had problems like that in past with git. Seems that is caused by editing file (you told me yesterday to change one line but it wasn't changed back)
[08:53:02] <Bonn1> I type RTAPI instead ?!
[08:53:34] <psha> git reset --hard
[08:54:30] <Bonn1> lot better but not yet ready
[08:54:32] <Bonn1> slavko@emc2:~/emc-camunits$ make
[08:54:32] <Bonn1> gcc -I. -Iemc2/ -fPIC -I/usr/local/include -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -I/usr/include/emc2 -DULAPI -c -o halio.o halio.c
[08:54:32] <Bonn1> Package glu was not found in the pkg-config search path.
[08:54:32] <Bonn1> Perhaps you should add the directory containing `glu.pc'
[08:54:32] <Bonn1> to the PKG_CONFIG_PATH environment variable
[08:54:33] <Bonn1> No package 'glu' found
[08:54:33] <Bonn1> gcc -L/home/slavko/emc2-new/lib -shared -o halio.so halio.o -lemchal -lglut
[08:54:33] <Bonn1> slavko@emc2:~/emc-camunits$
[08:56:33] <psha> libglut3-dev
[08:56:37] <psha> ah
[08:56:47] <psha> but compiled fine?
[08:56:53] <psha> then don't mind :)
[08:56:56] <psha> that's my fault
[08:57:35] <Bonn1> ??? I see error?!
[08:57:58] <Bonn1> ah I see halio.so
[08:58:03] <Bonn1> where to put?
[08:58:45] <psha> /usr/lib/camunits
[09:00:08] <Bonn1> please command line (in GUI there are not enought privilegies)
[09:01:31] <psha> sudo cp ...
[09:02:06] <Bonn1> :(
[09:02:12] <Bonn1> ok its done.
[09:03:00] <Bonn1> but here are no DRO readout still.
[09:03:51] <psha> axis is running?
[09:04:05] <Bonn1> EMC axis is running
[09:04:45] <psha> anything in stderr?
[09:04:57] <Bonn1> http://pastebin.com/GiPzS1qZ
[09:05:15] <Bonn1> I run EMC from terminal.
[09:06:46] <Bonn1> seems that only orientation is problem.
[09:06:55] <Bonn1> 'only'
[09:07:27] <psha> you have home brewed chain?
[09:07:34] <psha> or running camview-emc without any arguments?
[09:07:46] <Bonn1> home chain
[09:07:53] <psha> expand controls and see if there is emc.halio in list of tabs
[09:08:18] <Bonn1> not
[09:08:31] <psha> so add it with camview
[09:10:03] <Bonn1> ?? there are no such thing in camview!
[09:10:35] <Bonn1> read write data to HAL pins?
[09:12:16] <psha> yes
[09:14:28] <Bonn1> http://pastebin.com/6DXkQRYh
[09:15:02] <Bonn1> the tab is here, dro still not work and crosshair are not drawn (tab is present)
[09:15:17] <psha> it's ok that video.circle is not found
[09:16:40] <psha> both crosshair and halio tabs are present?
[09:16:52] <psha> in which order? crosshair, halio or halio, crosshair?
[09:17:51] <Bonn1> croshair,halio
[09:18:25] <Bonn1> .. main,extra,input,flip,croshair,halio,opengl
[09:19:17] <psha> opengl is overwriting everything :)
[09:19:24] <psha> flip,opengl,crosshair,halio
[09:19:34] <psha> crosshair and halio are overlays
[09:22:33] <Bonn1> ok I got image with DRO
[09:23:56] <Bonn1> but DRO has wrong numbers!
[09:24:02] <psha> :)
[09:24:10] <psha> fix axis pins in source
[09:24:14] <Bonn1> when G43 is on the numbers are wrong
[09:24:15] <psha> to correct one
[09:24:45] <Bonn1> when G43 is off (g49) then numbers are correct!
[09:25:09] <Bonn1> same as mine DRO yesterday!
[09:25:24] <psha> snprintf(name, 255, "halui.axis.%d.pos-feedback", i);
[09:25:35] <psha> http://psha.org.ru/cgit/psha/emc-camunits/tree/halio.c#n120
[09:25:56] <psha> put your name here
[09:26:53] <Bonn1> ???
[09:28:00] <Bonn1> I had wrong numbers until I subtract halui.tool.length_offset.x
[09:31:04] <psha> as i recall you've added 'add' component to hal so you have pins with correct names now?
[09:32:03] <Bonn1> yes that's true...
[09:32:56] <psha> so replace "halui.axis.%d.pos-feedback" with "you-pin-name.%d" if you pins are you-pin-name.0, you-pin-name.1, ...
[09:33:09] <psha> 0,1,.... are axis indices
[09:34:27] <Bonn1> How to do that. There are loop to pass over all axes. To destroy loop and only do 3x sprintf 1x for each axes?
[09:35:10] <Bonn1> ... no I haven't named pins like 0,1... but other.
[09:37:11] <Bonn1> and DRO output is wrong even if I do TouchOff
[09:37:15] <psha> if pin is not found then it's not initialized
[09:37:21] <psha> so if you have only 3 it's all right
[09:38:12] <Bonn1> hmm... that file I need to compile ?!
[09:38:15] <psha> just create 3 pins correct.0 correct.1 correct.2 for correct values and replace string with "correct.%d"
[09:38:37] <psha> after you've changed it run 'make' and copy halio.so to /usr/lib/camunits
[09:38:49] <psha> sudo make install will work as i recall
[09:39:55] <Bonn1> Hmm if I change source then I wan't to show axis names instead numbers (XYZ instead 012)
[09:43:16] <psha> it's a bit harder but also possible
[09:44:00] <psha> wait a bit, i'll add both adjustable axis format and letters
[09:44:11] <Bonn1> I don't want to have some custom file as if you change something then I need to recompile that
[09:45:37] <psha> i'll add pin names to parameters
[09:45:45] <psha> so you don't need to have you own version, just config
[09:45:46] <Bonn1> I got correct DRO with halui.axis.0.pos-relative (minus) halui.tool.length_offset.x
[09:46:37] <psha> XYZABCUYV
[09:46:47] <psha> if thees correct axis name sequence?
[09:47:06] <Bonn1> for me XYZ
[09:47:15] <Bonn1> don't know for others
[09:48:06] <Bonn1> halui.tool.length_offset.x are named by axis name not number.
[09:48:46] <Bonn1> I think (well I'm near sure) that all machines need that to do correct output
[09:49:21] <Bonn1> (without having custom calculation in HAL )
[09:49:36] <psha> that has to be fixed in halui, not in video
[09:50:19] <Bonn1> hmm the right names are declared in INI file
[09:50:44] <Bonn1> [TRAJ]
[09:50:55] <Bonn1> AXES= X Y Z ... whatever
[09:51:00] <psha> i've no way of reading INI file
[09:51:48] <Bonn1> Then just use some commandline switch to declare just right axes names for other than XYZABC...
[09:52:05] <Bonn1> I think lathe doesn't use XYZ
[09:53:09] <Bonn1> I think to change snprintf(name, 255, "halui.axis.%d.pos-feedback", i);
[09:53:36] <Bonn1> to use halui.axis.0.pos-relative
[09:53:44] <Bonn1> at least....
[09:54:41] <Bonn1> but if can then just do i=halui.axis.%d.pos-relative
[09:55:07] <Bonn1> i=i- halui.tool.length_offset.(pinname)
[09:55:34] <Bonn1> that's seems to be correct formula to have DRO compatible to axis one...
[09:55:39] <psha> ?
[09:55:58] <psha> i'll better add this to halui :)
[09:55:59] <Bonn1> I didn't check the source but in that way I get consistent reading from both
[09:56:45] <Bonn1> Is that hard?!?! (that seems easy to do for me)
[09:57:08] <Bonn1> and as I know the HAL component's disturbs RT performance
[09:57:15] <psha> it's not very hard but it's really not that place where to do it
[09:57:19] <Bonn1> so I don't like to have to much in HAL
[09:57:26] <psha> for example in my tasks i need only absolute position
[09:58:14] <Bonn1> ... but I have CNC and realy don't like to have DRO on one window to show different value/
[09:58:18] <psha> if it will be added into halui then it's not needed to do any HAL magick
[09:58:28] <Bonn1> ?!?
[09:58:55] <Bonn1> Em I miss something?
[09:59:27] <Bonn1> seem's that I mess term halui with something...
[10:01:17] <Bonn1> I need all that http://pastebin.com/xUH01ftd in my hal file to do correct DRO. It should be somewhere else I think. (but maybe I'm wrong)
[10:01:24] <psha> yes, you miss
[10:01:35] <psha> halui is userpsace component which provides you with a lot of useful pins
[10:01:55] <psha> for example halui.tool.length_offset.x
[10:02:13] <Bonn1> hmm I know that :D
[10:03:19] <Bonn1> So you intend to put some pins here like halui.DRO.axis.0 here?
[10:03:50] <Bonn1> that's probably change in EMC source?
[10:04:31] <psha> yes
[10:04:37] <psha> but if it's reasonable - why not
[10:05:03] <Bonn1> correct.
[10:05:23] <Bonn1> Just wonder what formula AXIS use to show DRO....
[10:06:41] <Bonn1> (I mean why isn't that done already as all UI have DRO?!?)
[10:08:21] <psha> git reset --hard; git pull; make
[10:08:24] <Bonn1> just discovered that order in halui for lenght offset is abcuvwxyz if that mean something
[10:08:59] <psha> default in video is 012345678 :)
[10:09:10] <psha> so place what you want :)
[10:10:19] <Bonn1> hmm numbers are now off a lot more...
[10:12:12] <Valen> * Valen has just purchased a chineese spindle
[10:12:20] <Valen> * Valen is nervous
[10:12:51] <psha> set correct pin template in emc.halio configuration
[10:13:04] <psha> for halui pins i have to consult other devs
[10:16:50] <Bonn1> something is wrong.
[10:16:57] <psha> ?
[10:17:27] <Bonn1> I set halui.axis.%d.pos-relative
[10:17:44] <Bonn1> but walue seems to be wrong one.
[10:18:21] <Bonn1> halui.axis.%d.pos-relative does manage touch off offsets but value on DRO doesn't
[10:19:45] <Bonn1> I run halmeter on halui.axis.1.pos-relative and i see value of 10 (correct) but dro shows 97
[10:20:07] <psha> have you set parameter in emc.halui settings?
[10:20:09] <Bonn1> =pos-commanded
[10:20:13] <Bonn1> yes
[10:20:33] <Bonn1> I do set,save, and restarted EMC!
[10:21:14] <psha> i've just tested and pos-relative shows me correct results
[10:22:32] <Bonn1> hmm... saving is problem.
[10:22:52] <Bonn1> The value is correct until I restart EMC
[10:23:13] <Bonn1> I do cahnge text and do SET.. The value is ok.
[10:23:27] <Bonn1> Then I save configuration..
[10:23:33] <Bonn1> and restart EMC
[10:23:58] <Bonn1> the value is wrong until I click SET again. (the text is saved ok)
[10:24:30] <psha> rebuild
[10:24:33] <psha> i've fixed it
[10:27:47] <Bonn1> you don't :(
[10:27:54] <Bonn1> stil same..
[10:28:02] <Bonn1> work until reboot
[10:28:58] <psha> ?
[10:29:53] <Bonn1> I must click SET button every time I reboot
[10:30:36] <psha> reboot emc or reboot camview?
[10:30:41] <psha> have you pulled last changes
[10:30:42] <psha> ?
[10:30:53] <psha> 1c45846 halio: Fix pin reload on control change
[10:31:03] <Bonn1> I reboot EMC
[10:31:10] <Bonn1> camview is OFF all the time
[10:34:41] <psha> i've tested with pos-relative, saved chain and then loaded from it again
[10:34:44] <psha> and it works
[10:36:48] <Bonn1> hmm.. after few times that change...
[10:37:03] <Bonn1> Now I get correct reading when boot up..
[10:37:27] <Bonn1> but value doesn't change when JOG machine until I click SET!
[10:39:15] <Bonn1> I reproduced that two times.
[10:39:33] <Bonn1> Start EMC the walue is correct.
[10:39:46] <Bonn1> jog machine the walue is frozzen
[10:40:06] <Bonn1> click SET the value change and folow JOG
[10:42:15] <psha> with emc2-sim it's working just fine...
[10:42:30] <psha> you've connected them to relative or to corrected pins?
[10:42:38] <Bonn1> relative
[10:43:11] <Bonn1> (relative from halui not from my HAL result)
[10:43:25] <psha> btw what is SET? is it 'Reloat hal pins' checkbox?
[10:43:37] <psha> or something else?
[10:43:42] <Bonn1> SET button
[10:44:49] <Bonn1> right to Base name for Axis pins
[10:45:12] <Bonn1> and I have halui.axis.%d.pos-relative there
[10:45:22] <Howard> psha, will you be about tomorrow ?
[10:45:33] <Bonn1> ?
[10:45:34] <psha> yes
[10:45:52] <Bonn1> sry to quick respond.
[10:45:54] <psha> Bonn1: ah, found
[10:46:22] <Howard> OK I'll be in contact about the same time, till tomorrow Bye.
[10:46:29] <psha> bb
[10:48:52] <Bonn1> psha can you limit number of decimal places and make little biger font too?
[10:50:07] <psha> decimal places - yes, via adjustable format string
[10:50:13] <psha> bigger font - dunno how :)
[10:50:38] <Bonny> is it ready to pull?
[10:53:35] <psha> yes
[10:55:06] <psha> Bonn1: for fonts there are only values HELVETICA 10 12 18
[10:57:09] <psha> i think i'll better switch to cairo
[10:57:12] <psha> but it's not easy :(
[10:59:54] <Bonn1> hmm still same problem.
[11:00:13] <Bonn1> value are frozzen until I click SET
[11:01:03] <Bonn1> The TAB DRO in axis save 'just right' font.
[11:01:21] <Bonn1> maybe just setting BOLD ?
[11:03:34] <Bonn1> I set Basename to halui.axis.%d.pos-feedback and value are not frozen
[11:03:50] <Bonn1> ie folow JOG instantly
[11:04:20] <Bonn1> but with halui.axis.%d.pos-relative I need to click SET every time I startup EMC
[11:22:03] <UncleG> Today is great day.
[11:26:55] <UncleG> EMC + Mesa 5i20 + 2 servos, and 24 air valves driving a index werke screw machine ready to run 2,200 lbs of 1" round 1215 steel....which will be unloaded off the semi 4 hours :) Thanks for everyone for helping me along; such as pcw_home SWPadnos Cradek and many more =D
[11:28:04] <UncleG> Not bad for a 23 year old HighSchool dropout =D
[11:28:10] <jthornton> take a video of it
[11:29:50] <UncleG> DEAL!
[11:30:02] <UncleG> I need a good camera, I have none :(
[11:30:06] <jthornton> what kind of part are you making
[11:30:37] <UncleG> Its just a round barrel with two different sized holes
[11:30:55] <UncleG> they are part of a heavy duty deadbolt for trash dumpsters
[11:31:34] <UncleG> I put an air driven recess tool in it so I wont have to second operation for a the bur in the inside rear hole
[11:32:04] <jthornton> cool
[11:33:52] <jthornton> I rarely make more than 2 of a part so I'm always changing over :/
[11:42:17] <UncleG> yeah they need 5,000 pcs
[11:42:20] <UncleG> O.o :)
[11:43:09] <UncleG> Maybe if I need some small runs done we can share work, whats your machinery?
[11:47:20] <jthornton> Hardinge CHNC chucker lathe, BP Discovery 308 VMC
[11:47:32] <jthornton> jthornton plasma :)
[11:53:22] <jthornton> maybe take a couple of photos of it and share on the forum or wiki
[11:53:31] <jthornton> sounds like an interesting machine
[13:32:03] <jthornton> looks like the rains have stopped finally and my levee has held up... this time
[13:37:54] <servos4ever> Does anyone have an example of how to use the on board LEDs ((bit in) CR<NN>) on the MESA 7i43 via HOSTMOT2 in HAL?
[13:39:02] <SWPadnos> do you see CR<NN> pins in HAL?
[13:39:34] <servos4ever> I'll look
[13:42:51] <SWPadnos> you should have those pins available. just hook them to some bit signal and the LED will turn on when the signal is true
[13:43:01] <SWPadnos> you can test it out by hand using halcmd:
[13:43:17] <SWPadnos> halcmd setp hm2_5i20.0.cr00 1
[13:43:19] <SWPadnos> halcmd setp hm2_5i20.0.cr00 0
[13:43:38] <SWPadnos> you should see the LED turn on after the first command and off after the second
[13:43:43] <SWPadnos> (the first LED that is)
[13:43:48] <servos4ever> No, but I'm not certain where they would appear. I don't see them under hm2_7i43 or below.
[13:44:06] <SWPadnos> oh. does the 7i43 have user-usable LEDs?
[13:44:35] <SWPadnos> oops - it looks like they start with CR01, not CR00
[13:46:16] <servos4ever> there are 8, and hostmot2 docs say they start with CR01. I just can't seem to get it...
[13:47:17] <SWPadnos> with emc running, open a terminal and run this command:
[13:47:26] <SWPadnos> halcmd show pin | grep CR
[13:47:52] <SWPadnos> if you get no output, then the LEDs aren't available to you (which is probably a bug)
[13:48:10] <SWPadnos> if you get a list of pins, then they are
[13:49:55] <mshaver> ON my setup I get hm2_5i20.0.led.CR01-8, just for reference... :) Of course, it should be 0-7 if we are true geeks!
[13:50:24] <SWPadnos> I think he might have matched the silkscreen :)
[13:54:19] <servos4ever> I tried "halcmd show pin | grep CR" in a terminal, but didn't see any output.
[13:55:29] <servos4ever> hostmot2 docs say the pins begin with CR01, and correspond to the PCB silkscreen
[13:56:31] <servos4ever> I would have assumed they would begin with 0 or 00 too, but the docs say 01
[13:58:26] <servos4ever> is "hm2_5i20.0.led.CR01-8" the whole line, or is there something else like newsig, setp, net, etc? - I'm new at this.
[14:07:43] <mshaver> Sorr, AFK for a minute! It's hm2_5i20.0.led.CR01, hm2_5i20.0.led.CR02, hm2_5i20.0.led.CR03, etc. This assumes you have a 5i20.
[14:08:40] <mshaver> I'm loading a kind of special bit file at the moment, but I think these are in the regular ones as well.
[14:09:08] <SWPadnos> servos4ever, with a 7i43, you would use hm2_7i43.0.led.CR01
[14:09:31] <SWPadnos> halcmd setp hm2_7i43.0.led.CR01 1
[14:09:38] <SWPadnos> should turn on the first LED
[14:09:43] <SWPadnos> halcmd setp hm2_7i43.0.led.CR01 0
[14:09:49] <SWPadnos> should turn it back off
[14:10:12] <SWPadnos> (assuming that emc is running, or you have the 7i43 driver loaded and the realtime system running some other way)
[14:11:23] <mshaver> Using the exact halcmd line you quoted (copy & paste exact) I get a display of these signal with firmware=hm2/5i20/SVST2_4_7I47.BIT
[14:12:17] <mshaver> Are these CR bits listed in the 7i43 .pin file?
[14:12:37] <servos4ever> here's the error I get if I try turning it on that way...
[14:12:40] <servos4ever> Invalid argument
[14:12:40] <servos4ever> Invalid argument
[14:12:40] <servos4ever> while executing
[14:12:40] <servos4ever> "hal halcmd setp hm2_7i43.0.led.CR01 1"
[14:12:40] <servos4ever> ("eval" body line 1)
[14:12:41] <servos4ever> invoked from within
[14:12:43] <servos4ever> "eval hal $what"
[14:12:45] <servos4ever> (procedure "showEx" line 3)
[14:12:47] <servos4ever> invoked from within
[14:12:48] <SWPadnos> they should be. I can't seem to find my 7i43 here
[14:12:49] <servos4ever> "showEx $halcommand"
[14:12:51] <servos4ever> invoked from within
[14:12:55] <servos4ever> ".main.note.fps.b.execute invoke"
[14:12:57] <servos4ever> ("uplevel" body line 1)
[14:12:59] <servos4ever> invoked from within
[14:13:01] <servos4ever> "uplevel #0 [list $w invoke]"
[14:13:03] <servos4ever> (procedure "tk::ButtonUp" line 22)
[14:13:05] <servos4ever> invoked from within
[14:13:07] <servos4ever> "tk::ButtonUp .main.note.fps.b.execute"
[14:13:09] <SWPadnos> the lines I gave you are tho be run from a terminal, not from halshow
[14:13:09] <servos4ever> (command bound to event)
[14:13:37] <servos4ever> I have EMC running, and I'm using the show HAL configuration tool.
[14:13:44] <SWPadnos> yes. don't do that :)
[14:14:03] <SWPadnos> the command line is more flexible, and is what I have been giving examples for
[14:14:16] <servos4ever> they are _not_ listed in the pin file
[14:14:28] <SWPadnos> ok, in that case there's something wrong.
[14:14:37] <mshaver> just start a second terminal and paste them in like: mshaver@Matt:~$ halcmd show pin | grep CR
[14:14:37] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR01
[14:14:37] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR02
[14:14:37] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR03
[14:14:37] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR04
[14:14:38] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR05
[14:14:38] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR06
[14:14:39] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR07
[14:14:39] <mshaver> 9 bit IN FALSE hm2_5i20.0.led.CR08
[14:14:40] <mshaver> mshaver@Matt:~$
[14:15:02] <SWPadnos> I can't find my 7i43 at the moment (I'm sure it's here somewhere), so I can't try to figure out what's wrong
[14:15:10] <servos4ever> all the expected IO pins are in the pin file
[14:15:59] <servos4ever> but no CRxx or LEDxx are there
[14:16:32] <mshaver> well, if the firmware is loading, they should be there. Do your other pins, like GPIOs show up?
[14:17:24] <servos4ever> all the "other" pins are there
[14:17:54] <pcw_home> What Version EMC? LED access in the driver is pretty recent
[14:18:08] <mshaver> Well I'll be monkey's uncle!
[14:19:08] <servos4ever> there are other things that don't appear in the pin file, like .is_output, .is_opendrain, etc.., so i was not sure if they should appear or not.
[14:20:35] <SWPadnos> those are parameters, not pins :)
[14:23:20] <servos4ever> version 2.4.5, upgraded from 2.3.? using the instructions on WIKI
[14:23:59] <servos4ever> see what I mean about being new to all this ;)
[14:26:06] <mshaver> I'm running a compiled version of 'master' in the git repository (2.5.0~pre), but I think my firmware is from 2.4.4 or 5, so you should be PK.
[14:26:32] <mshaver> OK, not PK...
[14:26:49] <mshaver> Well, maybe you're PK too!
[14:27:36] <servos4ever> I probably am. I feel kinda PK...
[14:27:53] <SWPadnos> psychokinetic?
[14:27:55] <mshaver> You could run your hal script and the pastebin the output of dmesg for Peter to look at.
[14:28:16] <mshaver> U yea, psycho something!
[14:28:41] <mshaver> I meant, "Uh"... It's early yet.
[14:29:16] <SWPadnos> psycho killer, qu'est-ce que c'est?
[14:30:15] <pcw_home> Dont think its a firmware issue, its more likely a driver issue (Are LEDS only supported in Master?)
[14:31:09] <servos4ever> my HAL script runs just fine - until I try to add anything about LEDs. Except..., I do have one error message I been ignoring.
[14:31:33] <SWPadnos> pcw_home, that's possible
[14:32:07] <mshaver> You know, I've never noticed them before installing master to do the 8i20 stuff...
[14:32:14] <morficmobile> ah, pcw_home, reminds me, i need to get a quote this week
[14:32:39] <servos4ever> I always get a parport mode 4 not supported error - "continuing anyway" every time I run Axis / EMC2
[14:32:45] <SWPadnos> yeah, it looks like Andy P added support in May of this year, but the change was never backported to 2.4
[14:32:55] <mshaver> How about, "To be or not to be...", that's a good quote.
[14:33:17] <SWPadnos> how about "no matter where you go, there you are"
[14:33:32] <mshaver> I've also always like, "No good deed goes unpunished." - Dorothy Parker
[14:33:43] <SWPadnos> heh
[14:34:04] <pcw_home> 0x2B | ! 0x2B
[14:34:07] <servos4ever> not being there - that might explain why I can't figure it out...
[14:34:18] <jthornton> "Don't push THAT button!"
[14:34:33] <SWPadnos> servos4ever, it does make it more difficult
[14:34:35] <mshaver> Is that Yogi Berra?
[14:34:47] <mshaver> I think it is.
[14:35:01] <SWPadnos> Buckaroo Banzai
[14:35:47] <servos4ever> should I just forget about LEDs for now...?
[14:35:53] <mshaver> I sit corrected.
[14:36:11] <mshaver> you could install master
[14:36:16] <SWPadnos> servos4ever, there are several ways to run Master instead of 2.4.x, you can do any of those, or forget it for now :)
[14:36:23] <servos4ever> I really like blinky lights, but I can live without them
[14:36:59] <pcw_home> You can use raw-write if you must have blinky-lights
[14:37:00] <SWPadnos> the Installing EMC2 page tells how to compile the source, and I believe there are experimental packages that are automatically built whenever someone makes a change
[14:37:06] <servos4ever> the more blinky lights and meters the better I like it, but maybe I should move on.
[14:37:11] <SWPadnos> (Installing EMC2 on the wiki, that is)
[14:37:25] <mshaver> does the buildbot still make them (nightlys)?
[14:37:35] <SWPadnos> at every checkin, I think
[14:37:48] <SWPadnos> I'm just not exactly sure where to get the packages
[14:38:28] <mshaver> seb would know
[14:39:30] <servos4ever> you guys know how "feature creep" sets in, and then...
[14:39:54] <servos4ever> "the task at hand expands to fill the time available"
[14:40:47] <SWPadnos> yes. tasks are like liquids that way
[14:40:58] <servos4ever> more like gasses
[14:40:59] <SWPadnos> err, fluids
[14:41:12] <SWPadnos> gases and liquids are both fluids :)
[14:42:51] <servos4ever> right you are, but I meant expanding in a 3D space rather than 2D
[14:43:12] <SWPadnos> or 4d, in the case of tasks :)
[14:43:33] <servos4ever> roger that!
[14:49:51] <servos4ever> Thanks for your help everyone. I'll skip the LEDs for now since I don't need them anyway. I'll try again after the next release.
[14:52:27] <servos4ever> I was hoping to bring all this MESA hardware to Stuart's shop a few weeks ago, but just could not get away.
[14:55:11] <servos4ever> After this 7i43, 7i37, 7i33 project i'm planning to use a 5i20 with a 7i48, 7i37, etc. for a 5 axis CNC mill.
[14:56:02] <pcw_home> 7I48 needs master as well... (or maybe 2.5 will come out)
[14:56:18] <servos4ever> Will I need to use something other than the current release for the 7i48?
[14:56:44] <servos4ever> You answered before I could ask! Thanks.
[14:58:44] <servos4ever> SWPadnos - I just used the last one of those 24vdc power supplies I bought from you in about 2006. They are still under warranty, right?
[14:59:07] <SWPadnos> heh, I wonder ... :)
[14:59:16] <SWPadnos> do they all work?
[14:59:29] <servos4ever> Yea, they're perfect.
[14:59:35] <SWPadnos> (if so, then they're under warrantee :) )
[14:59:45] <SWPadnos> ok. the lifetime warrantee ends when they die ...
[15:01:32] <servos4ever> One more real question before I leave.
[15:03:24] <servos4ever> Is there any way to get an "analog" signal into EMC2 using the 7i43, 7i33, 7i37 combination?
[15:03:45] <cradek> what are you trying to do?
[15:04:12] <SWPadnos> if you can make a voltage to frequency converter, then you can get a pseudo-analog input by connecting that to an encoder input in counter mode
[15:07:25] <jthornton> Mesa THC is a voltage to frequency converter
[15:08:39] <servos4ever> I really want a some kind of functional "feed override POT".
[15:08:53] <cradek> aha, that's exactly why I asked
[15:09:07] <SWPadnos> use an encoder. pots are not recommended
[15:09:07] <cradek> the way to do that with EMC is using an encoder knob hooked to halui
[15:09:30] <cradek> each click of the encoder is so many percent change in feed override
[15:10:09] <servos4ever> can I use GPIO for this? I'm using 4 encoders as encoders already
[15:10:54] <jthornton> I hooked up a pokeys with 4 pots for FO etc. and tested it out but never did install it
[15:11:03] <cradek> yes, you can just turn on more encoder counters, and hook up to the right input pins
[15:11:42] <servos4ever> ...or could you point me to an example on the wiki or something? I thought I could ony have 4 encoders with this hardware combination
[15:12:02] <cradek> turn on more of them - check dmesg to see what pins they show up on - wire accordingly
[15:12:08] <SWPadnos> check the firmware options for the 7i43. there may be one with 8 "servo channels"
[15:12:36] <cradek> what firmware are you using?
[15:14:31] <servos4ever> Nov 30 08:39:53 dah-lih-desktop kernel: [ 4727.863735] hm2: loading Mesa HostMot2 driver version 0.15
[15:14:31] <servos4ever> Nov 30 08:39:53 dah-lih-desktop kernel: [ 4727.870098] hm2_7i43: loading HostMot2 Mesa 7i43 driver version 0.3
[15:14:31] <servos4ever> Nov 30 08:39:53 dah-lih-desktop kernel: [ 4727.870108] PARPORT: linux parport parport0 does not support mode 4.
[15:14:31] <servos4ever> Nov 30 08:39:53 dah-lih-desktop kernel: [ 4727.870109] PARPORT: continuing anyway.
[15:16:11] <SWPadnos> you should be able to use the ..SV8 version of the firmware, unless you need some stepgens
[15:17:41] <servos4ever> Also, I've tried different BIOS settings to eliminate the PARPORT error, but no luck. I'm pretty sure it was not there with 2.3
[15:18:02] <SWPadnos> if you started with one of the sample configs, then you can change the [HOSTMOT2]CONFIG line in your ini file
[15:18:07] <cradek> just ignore it, if the device works. it's a (mis)feature of your system's bios.
[15:18:16] <SWPadnos> change this: CONFIG="firmware=hm2/7i43/SVST4_4B.BIT num_encoders=3 num_pwmgens=3 num_stepgens=0"
[15:18:19] <SWPadnos> to this:
[15:18:29] <cradek> EPP is the correct bios setting
[15:18:31] <SWPadnos> CONFIG="firmware=hm2/7i43/SV8.BIT num_encoders=3 num_pwmgens=3 num_stepgens=0"
[15:18:33] <servos4ever> I don't need any stepgens if they are only needed for steppers.
[15:18:57] <SWPadnos> err, well without the stepgens word, and changing the number of PWM/encoders to suit your machine
[15:19:13] <SWPadnos> CONFIG="firmware=hm2/7i43/SV8.BIT num_encoders=5 num_pwmgens=4"
[15:19:16] <SWPadnos> or whatever
[15:19:49] <SWPadnos> the main thing is changing SVST4_4 to SV8, and you'll need a B or S on there, depending on which FPGA size you have
[15:20:23] <servos4ever> I tried ECP and EPP both just to make sure, but I still get the error. It was not there until I upgraded to 2.4.5
[15:20:54] <SWPadnos> 2.4 deals with parallel ports differently than 2.3 did
[15:22:15] <servos4ever> Dell GX280 box
[15:24:21] <servos4ever> ...also, the latency test results went from ~6k to about ~27k when I went to 2.4.5. Is this related to the parport error?
[15:25:02] <cradek> updating emc does not affect those numbers
[15:25:08] <SWPadnos> hmmm. probably not, but that doesn't sound good
[15:25:09] <cradek> did you change something else too?
[15:25:20] <cradek> 27k is perfectly fine for a servo machine
[15:26:39] <servos4ever> No, no other changes at all.
[15:26:57] <servos4ever> I'll explain better...
[15:30:05] <servos4ever> I installed the 10.04 & 2.4.5 from the live CD when it was released. I noticed the latency was much worse than with the previous 8.04 & 2.3? live CD, so I did another install with my old CD - and the latency was much better at about ~6-7K.
[15:30:20] <cradek> so you changed the entire OS, not just emc
[15:30:42] <cradek> it's definitely possible that affects it
[15:30:57] <cradek> still, 27k is perfectly great for your mesa-based servo system
[15:31:19] <servos4ever> I thought something in Ubuntu 10.04 may have caused this, so I installed 8.04 and upgraded EMC to 2.4.5 by the instructions on the wiki.
[15:31:40] <cradek> ok, that's fine too
[15:32:08] <cradek> the difference is likely that you got a different video driver
[15:34:25] <JT-Hardinge> how would you get a different video driver when upgrading from 2.3 to 2.4?
[15:34:40] <cradek> JT-Hardinge: reading comprehension failure!
[15:34:53] <JT-Hardinge> I suffer from that as well
[15:34:57] <servos4ever> So, Ubuntu 8.04 & EMC 2.3? latency was ~6-7k, and 8.04 & EMC 2.4.5 latency is about ~27k.
[15:35:20] <cradek> oh wait, it's me with the failure
[15:35:52] <cradek> emc 2.3 vs emc 2.4 will not affect your latency numbers. the latency test is exactly the same. something else is different.
[15:35:58] <servos4ever> The failure is probably the way I'm telling the story...
[15:36:08] <cradek> although, it's moot, since as I've said several times, both of those latency results are perfectly fine
[15:37:45] <SWPadnos> have there been any RTAI kernel updates to 8.04?
[15:37:55] <cradek> I don't think so
[15:38:06] <servos4ever> If the ~27k latency is not a problem, then I don't have a problem at all. I don't have the servos going yet to test, and this is my first experience with the MESA hardware and EMC together.
[15:38:35] <cradek> 27k is good enough even for software stepgen. you will have no trouble.
[15:38:50] <SWPadnos> is the latency test actually the same? I thought there was an option to change the base and servo periods now (I don't know when it was added)
[15:38:54] <cradek> a servo machine needs updates only every ms
[15:39:13] <servos4ever> Great - that's what I was hoping to hear.
[15:39:33] <SWPadnos> it's possible that a change in the base thread period would change the latency, due to cache effects
[15:39:38] <SWPadnos> (or something else unfathomable)
[15:42:18] <servos4ever> Back to the hostmot firmware and more encoders topic - would I need something other than a 7i37 connected to the P3 connector of the 7i43 in order to use any additional encoders that may appear in firmware?
[15:43:02] <cradek> no - you can use the 7i37 inputs, but beware they are pretty slow. there may be a maximum speed you can turn the knob, and it will stop responding if you turn it faster.
[15:43:42] <servos4ever> O.K. - thanks!
[15:43:43] <cradek> if you have trouble, you could peel those wires off before the 7i37.
[15:46:07] <servos4ever> would I need to jumper the 7i43 for 5v tolerant mode to use 2 TTL level encoder signals on the inputs ?
[15:46:47] <cradek> that I don't know
[15:49:31] <servos4ever> the 7i43 docs warn of signal integrity issues due to pull up resistors vs: 5v tolerant mode and TTL level inputs, but I cannot fully understand what it means.
[15:55:24] <servos4ever> Thanks to all for your help so far. I was feeling pretty alone for a while, but I'm O.K. (or PK) for now.
[15:55:44] <servos4ever> and thanks again for an AWESOME package!
[15:56:40] <servos4ever> You guys are my heros!
[16:06:39] <Bonny> psha are you ok?
[16:06:44] <servos4ever> I just rebooted and confirmed bios for onboard parport is indeed set for EPP mode, but I still get the error message when I launch EMC. Is this O.K.?
[16:07:03] <SWPadnos> servos4ever, yes, it's fine as long as the 7i43 is working
[16:07:34] <SWPadnos> the message says "Linux told me it might not work, but I'll assume it does anyway and just keep on going"
[16:07:46] <servos4ever> Nov 30 07:50:25 dah-lih-desktop kernel: [ 1764.902772] hm2: loading Mesa HostMot2 driver version 0.15
[16:07:46] <servos4ever> Nov 30 07:50:25 dah-lih-desktop kernel: [ 1764.907429] hm2_7i43: loading HostMot2 Mesa 7i43 driver version 0.3
[16:07:46] <servos4ever> Nov 30 07:50:25 dah-lih-desktop kernel: [ 1764.907438] PARPORT: linux parport parport0 does not support mode 4.
[16:07:46] <servos4ever> Nov 30 07:50:25 dah-lih-desktop kernel: [ 1764.907440] PARPORT: continuing anyway.
[16:08:12] <psha> Bonny: i'm via GPRS :(
[16:08:27] <Bonny> To much snow?
[16:08:30] <psha> my provider broke my link
[16:08:36] <psha> to much idiots
[16:08:40] <Bonny> :D
[16:08:54] <psha> it seem that they forgot to plug my cable back
[16:08:54] <SWPadnos> right. Linux didn't detect the port capabilities correctly (or the BIOS / ACPI tables were wrong), but the EMC parport driver decided to go on anyway
[16:09:21] <Bonny> Ok then I will not bother you today..
[16:09:42] <psha> :) i won't be able to look at images/screenshots anyway...
[16:10:07] <Bonny> Just dro doesn't work as should...
[16:10:23] <psha> not tracking changes on pins?
[16:10:29] <Bonny> I make hal pins and configure into camview..
[16:11:21] <Bonny> and funny is that when I start EMC I don't see DRO at all. But when click that button right of the config text then start to work and is correct.
[16:11:26] <psha> may you test same setup on 10.04?
[16:11:57] <psha> so to track down if it's halio plugin bug only or it's triggered by hardy
[16:12:02] <Bonn1> probably not as I don't have same setup in 10.04
[16:12:26] <Bonn1> I think is problem with defaults.
[16:12:47] <servos4ever> This behavior is not present with the 8.04 & 2.3.x combination. It appeared when I upgraded EMC to 2.4.5. It does not seem to cause any real trouble, but I thought it might help someone to know this.
[16:12:48] <psha> so it's tracking pin changes? not only after pressing 'set' button?
[16:13:13] <Bonn1> pin sure change as I have labels connected to it and work
[16:14:14] <psha> DRO values connected to pins
[16:14:33] <Bonn1> sum2.%d.out is now base name for hal pins and does
[16:15:02] <Bonn1> and I have connected that pin to the label (my old way DRO) and it works
[16:15:10] <psha> DRO values are freezed and only changed after you press 'set'?
[16:15:23] <psha> and then are frozen again?
[16:15:30] <Bonn1> no
[16:15:47] <psha> DRO values are not present at all until you press 'set'?
[16:15:50] <Bonn1> with that new config I do Get:
[16:15:55] <Bonn1> yes !!!
[16:16:15] <psha> but correct values are filled in textbox?
[16:16:42] <Bonn1> In older one I got DRO value visible but are static (ie not changed) until I press SET
[16:16:57] <Bonn1> Yes now are all walues correct!
[16:17:21] <Bonn1> yes the box preserve my entered string
[16:17:35] <Bonn1> just need to click SET
[16:17:53] <psha> so only issue is that until you click Set first time it's missing?
[16:18:05] <Bonn1> seems like that
[16:18:18] <psha> in which file you hal setup is living?
[16:18:43] <Bonn1> the -H executed one.
[16:18:50] <Bonn1> (with camview)
[16:18:53] <psha> is it called before or after camview-emc start?
[16:18:54] <psha> ah
[16:18:54] <psha> that's why it's not working
[16:18:54] <psha> wait a bit
[16:19:27] <psha> theese pins are created after chain is loaded
[16:19:43] <Bonn1> I think that's problem.
[16:20:30] <Bonn1> as I don't see pin at all if pin is my one. If I get some pin from halui it's visible.
[16:20:37] <psha> if you move base part (calculating offsets) to earlier stage it'll work
[16:21:09] <psha> leave only 'net signame camview.pin' clauses into camview HAL file
[16:21:31] <psha> unfortunately i won't be able to talk about halui changes today :(
[16:21:44] <psha> so live for a while with this setup
[16:24:14] <Bonn1> I put that pins in main HAL file. Now I see DRO at startup but it's frozen until click SET
[16:24:34] <psha> ah! i know why :)
[16:24:49] <psha> it finds pin and then it's connected to signal
[16:25:46] <psha> give me a line from camview HAL file
[16:25:49] <Bonn1> but if the pin is -feedback from halui then works.
[16:25:51] <psha> for a single axis
[16:26:12] <Bonn1> what line?
[16:26:19] <psha> reason is that i'm cheating there :) i'm not using 'true' way and not creating new pins
[16:26:20] <psha> 'net' line
[16:27:18] <psha> for one axis, others are equals i think
[16:27:36] <Bonn1> loadrt sum2 count=3
[16:27:36] <Bonn1> addf sum2.0 servo-thread
[16:27:46] <Bonn1> setp sum2.0.gain1 -1
[16:27:49] <psha> no 'net' commands?
[16:27:55] <Bonn1> net sum20in0 halui.axis.0.pos-relative => sum2.0.in0
[16:28:08] <Bonn1> net sum20in1 halui.tool.length_offset.x => sum2.0.in1
[16:28:36] <Bonn1> actualy no net from output of SUM2 module!
[16:28:49] <Bonn1> net sum20out sum2.0.out => camview.hal_label1
[16:28:56] <Bonn1> but here for label is!
[16:29:18] <psha> there is not 'net' at all?
[16:29:19] <Bonn1> (different thing)
[16:29:32] <psha> or net is in camview HAL file?
[16:29:35] <psha> i mean sum2.0.out net
[16:29:40] <Bonn1> no net
[16:29:48] <Bonn1> .. from .out
[16:30:06] <psha> try this:
[16:30:48] <psha> net sum20out sum2.0.out
[16:30:51] <psha> so you create signal for that pin but don't connect anything
[16:33:22] <Bonn1> yes that work (I had added for all 3 axis)
[16:33:28] <psha> nice
[16:33:45] <Bonn1> now working
[16:33:59] <Bonn1> can that DRO have BOLD fonts?
[16:34:03] <psha> it seem that i've to create pins in camview-emc and in halio just poll them...
[16:34:39] <psha> heh, you really think that i know GLUT better then you :D
[16:34:46] <psha> i think yes but i don't know :)
[16:35:04] <psha> i'll check docs when my link will be up again
[16:35:08] <Bonn1> I have few thing in 2 hal files now to work.. :D
[16:35:10] <psha> and now i've to disconnect my phone
[16:35:20] <Bonn1> Ok have nice day
[16:35:24] <Bonn1> (night)
[16:35:26] <psha> :)
[16:35:33] <Bonn1> bb
[16:36:02] <psha> btw your suggestion about custom button to store current positions in file are easily achievable with recent mhaberler's changes to gladevcp
[16:36:17] <psha> so let look on this tomorrow
[17:30:32] <cradek> https://www.surpluscenter.com/item.asp?item=1-3431
[17:31:24] <Jymmm> thats cheap
[17:31:59] <cradek> yeah - if you need short ballscrews
[17:32:20] <SWPadnos> it
[17:32:25] <cradek> IT
[17:32:31] <cradek> looks like one circuit only
[17:32:33] <SWPadnos> it's big enough for a Bridgeport Y axis, more or less
[17:32:34] <Jymmm> With the finished ends, you could use them for anything
[17:33:34] <SWPadnos> hmmm. and definitely big enough to lift the ram on my bandsaw
[17:33:38] <Jymmm> Mind you, I think everyone uses smaller diameter becasue it's cheaper. and lighter but that doens't mean you couldn't use them ofr other projects.
[17:33:52] <cradek> yeah 1"-4 is pretty darn big
[17:34:45] <cradek> hm, they say acme - that must just be a mistake
[17:34:58] <SWPadnos> yeah, I was just noticing that
[17:35:12] <Jymmm> looks ACME to me https://www.surpluscenter.com/adphotos/P1-3431X1.jpg
[17:35:41] <SWPadnos> no, that's ball
[17:35:48] <SWPadnos> acme has corners in the grooves
[17:35:54] <SWPadnos> those are round-ish
[17:35:58] <Jymmm> ah
[17:36:08] <cradek> hard to tell with all the grease.
[17:36:10] <Jymmm> hard to tel withthe grease =)
[17:36:21] <SWPadnos> rolled ball, the outer diameter isn't shiny
[17:36:24] <cradek> I don't think you could really get balls to roll in acme
[17:36:39] <Jymmm> cradek: Sure you can.. once.
[17:36:41] <SWPadnos> well, I suppose you could, but why bother
[17:37:58] <Jymmm> oh crap I forgot about motor/pulley mount =(
[17:39:06] <SWPadnos> holy crap. that price is for a pair!
[17:39:46] <cradek> they have 964 pairs...
[17:40:03] <Jymmm> SWPadnos: Order me two pair, finish one end for motor/pulley mount and send em to me =)
[17:40:08] <SWPadnos> yeah, hence the $15 price :)
[17:40:13] <cradek> right.
[17:40:15] <SWPadnos> Jymmm, I'll get right on that :)
[17:40:33] <Jymmm> SWPadnos: lying bastard! ;)
[17:40:55] <SWPadnos> it's just behind the bandsaw modifications, which are behind my two larger retrofits, which are now behind my two smaller retrofits
[17:41:08] <cradek> wow, that's a lot of behindness
[17:41:11] <SWPadnos> and all that comes after work and play, so don't hold your breath ;)
[17:41:12] <Jymmm> who has a working cnc lathe?
[17:41:15] <SWPadnos> yes, I have a big behind
[17:41:16] <cradek> nobody
[17:41:22] <Jymmm> really?
[17:41:27] <cradek> yeah, there aren't any
[17:41:30] <SWPadnos> cradek, what's wrong with yours?
[17:41:38] <cradek> SWPadnos: (sshhhhhh)
[17:41:41] <SWPadnos> oops
[17:41:49] <SWPadnos> I mean, when do you expect to get yours working again?
[17:41:54] <SWPadnos> err
[17:41:54] <Jymmm> cradek: :(
[17:41:57] <cradek> Jymmm: aside from the fact that there aren't any, why do you ask?
[17:42:44] <Jymmm> cradek: Thos are cheap enough to start a new project, but I have no way of finishing the ends
[17:43:03] <Jymmm> Like I need another project....
[17:44:21] <Vq_> Vq_ is now known as Vq
[17:45:31] <Jymmm> SWPadnos: I got around the air scrubber vibrating the laser. I grabbed a mini dolly from HF and mounted the air scrubbers on wheels insted of on the laser itself. It still fits under the laser, just rolls independantly from it.
[17:46:01] <Jymmm> I just got done mounting the wheels on the scrubber yesterday.
[17:46:10] <SWPadnos> cool
[17:47:47] <Jymmm> SWPadnos: And I found out how to put 100% load on a DC with no output... run the air through carbon pellets.
[17:48:07] <SWPadnos> with no output?
[17:48:07] <Jymmm> *sigh*
[17:48:20] <Jymmm> Yeah, they block the air almost completely.
[17:48:22] <SWPadnos> you mean you can't feel any air coming out
[17:48:25] <SWPadnos> ah
[17:49:18] <Jymmm> I have to redesign the flow pattern to allow for more surface area. I *though* I did enough, but I was wrong.
[17:50:29] <Jymmm> I was trying for close to 100% contact, and was just too much for the DC.
[17:50:32] <SWPadnos> how big are the pellets? (and how are they in the air flow path - between cloth/filter paper, in a bucket ...)?
[17:50:45] <Jymmm> 4mm
[17:51:16] <Jymmm> divided into two stacked (and sealed) cardbaord boxes (for easy disposal).
[17:51:33] <SWPadnos> what thickness of pellets does the air travel through?
[17:51:35] <atmega> we have chip/fines collection systems that use water filters
[17:52:28] <Jymmm> SWPadnos: 5"
[17:52:53] <Jymmm> maybe 6" at points.
[17:52:57] <SWPadnos> that sounds awfully thick to me
[17:53:09] <Jymmm> Yeah, I know that *NOW* =)
[17:53:12] <SWPadnos> heh
[17:53:48] <Jymmm> I have more cardbaord boxes, so I'll increase the surface area to allow for better flow.
[17:54:23] <Jymmm> right now it's over 1cf of carbon. I'll reduce that by half and allow for better flow
[17:54:53] <Jymmm> I honsestly didn't think that the pellets would block so much air.
[17:55:32] <SWPadnos> well, remember that if they're round-ish, they pack very tightly together, and in 3 dimensions
[17:56:10] <Jymmm> Oh NOW he tells me ;)
[17:56:33] <SWPadnos> if they are pretty round, you could probably get good filter action by carefully laying out 2 or 3 orderly layers of pellets
[17:56:57] <Jymmm> Sidenote... My MBP video (int and ext) died the other day.
[17:57:12] <SWPadnos> put those between filter paper of some sort (something that's held somewhat rigid, like the aluminum stuff they use in furnace filters)
[17:57:25] <SWPadnos> AppleCare
[17:57:35] <Jymmm> SWPadnos: WAY out of waranty.
[17:57:46] <atmega> I have stainless screen at teh top/bottom of my co2 scrubber
[17:57:47] <SWPadnos> unless you don't have that, in which case you should just go buy two PCs to replace it (one plus a spare)
[17:58:11] <Jymmm> But Apple said that for a flat fee of $310 they will fix EVERYTHING and warrant the repairs for 90 days
[17:58:20] <SWPadnos> oh. that's not all that bad
[17:58:39] <SWPadnos> only half the price of a new (non-Mac) laptop
[17:58:57] <Jymmm> I didn't think so, I already got a new power supply, they'll fix the fans, video, and the dead dvd drive as well as replace the logic board
[17:59:21] <JT-Hardinge> if you want to install the sim only in Ubuntu 10.10 do these still apply? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Pure_Simulator
[18:00:00] <Jymmm> SWPadnos: Yeah, I could still shell, remote, and target mode into it, just no video =)
[18:01:33] <Jymmm> SWPadnos: I tossed it into target mode, and booted from it on the macmini.
[18:04:02] <Jymmm> SWPadnos: I duped the hdd to an ext usb hdd, and booted the macmini from that. then zeroed the drive before turning it in for repair yesterday.
[18:04:30] <Jymmm> It's been 20+ years, I wish PC's had that functionality
[18:04:40] <JT-Hardinge> or is this a better set of instructions on building a sim? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Building_emc2_simulator
[18:06:56] <JT-Hardinge> * JT-Hardinge finished his bowl of gruel and heads in to take a nap
[18:08:52] <Jymmm> JT-Hardinge: Buenos Noches
[18:09:04] <dimas> i'm getting error while building opencamlib, anybody could help here or in #cam?
[18:09:29] <dimas> bullcutter.cpp:135: error: aggregate ?std::ostringstream o? has incomplete type and cannot be defined
[18:09:43] <dimas> the same for other cutters
[19:12:08] <juri_> * juri_ read the integrator manual last night, for homebuilt-emc-machine ideas.
[19:13:48] <juri_> I'm trying to build mine for PCB fab.
[19:15:36] <psha> juri_: here is at least one user who is processing pcb's somehow
[19:19:14] <juri_> psha: whats your software stack look like? mine is EMACS->MUCS PCB->EMC2(hopefully).
[19:20:32] <psha> it's not me :)
[19:22:19] <skunkworks> eagle -> Gcode.ulp -> emc2
[19:23:39] <psha> skunkworks: hey, you are using you Huge one to drill PCB's?!
[19:24:34] <juri_> working out the math, my mill should have an X and Y stepping of ~8 microns, but i expect lots of tolerance-noise, due to this all being homebuilt. think thats good enough for PCB work?
[19:27:18] <psha> andypugh: good evening
[19:27:28] <andypugh> Hi there.
[19:27:45] <psha> i've some questions about your suggestion of finding center from 3 points
[19:28:19] <andypugh> OK
[19:29:24] <andypugh> I got the idea from reading this page years ago: http://miketreth.mistral.co.uk/centrecam.htm
[19:30:29] <psha> so workflow is following: you jog mill to first point, press some Magic Button (tm), then do it two more times and then get 3 points recorded
[19:30:29] <psha> but how you see process of removing bad points?
[19:30:29] <psha> question is not how to implement it, but how to put it in the interface :)
[19:30:29] <psha> implementation is trivial...
[19:31:33] <andypugh> Buttons that latch-down? Unlatch to try again. Then a "calc" button that produces the answer and releases them all?
[19:31:44] <psha> i've seen this prog already :) so it seem that it's good source of nice design concepts :)
[19:32:45] <andypugh> Yes, though I feel a bit guilty about ripping off too many of his ideas.
[19:33:06] <andypugh> Inverse of this: http://www.xkcd.com/
[19:34:10] <psha> andypugh: i'll try to draw some interface and show you how it looks like...
[19:34:35] <psha> if it's limited to 3 points it won't be difficult i hope
[19:34:51] <psha> or maybe 4 points are needed to increase precision?
[19:37:53] <andypugh> Won't hurt. However 3 points could also be used to measure angles. The interface could tell you circle diameter and centre, but also A-B-C angle too.
[19:38:54] <psha> hm, and to draw that on live view? :)
[19:46:56] <andypugh> If you want a challenge, feature tracking code could watch the image scroll past and make the points move with the work :-)
[19:47:49] <psha> heh, why to bother if you may do calibration and then calculate image coordinate from mill's :)
[19:48:44] <andypugh> Doesn't that calibration depend crucially on camera-to-work distance?
[19:49:16] <psha> certanly depends, but it's more easy then movement detection :)
[19:49:33] <andypugh> I think you can reliably measure positions by jogging the crosshair to them, but picking on-screen I can see being tricky.
[19:49:43] <cradek> isn't it easier to draw a circle on the video and let the user center it?
[19:50:03] <psha> cradek: it's already done :)
[19:50:05] <cradek> seems like a human would be better at lining up edges than picking points that are exactly on an edge
[19:50:33] <cradek> I'm saying I think the three point circle is not useful
[19:50:52] <psha> but if it's bigger then point of view?
[19:51:39] <andypugh> Yes, the three-point thing is for picking up big features like bearing bores.
[19:51:53] <cradek> a better way to find the center is to find the center in X, then go there, then find the center in Y, and repeat until it converges (it converges very fast)
[19:52:34] <andypugh> Nothing to stop you doing that if you prefer :-)
[19:53:11] <alex_joni> yay, got a new battery for my laptop
[19:53:12] <andypugh> (And surely it converges in 6 points?
[19:53:15] <cradek> 3 points is just a parlor trick IMO
[19:53:37] <andypugh> How about if you don't have a full circle?
[19:53:48] <cradek> hm, good question
[19:54:02] <cradek> I have not run into that situation
[19:54:45] <alex_joni> 4:26 hours (86%) remaining
[19:54:45] <psha> cradek: i suspect that additional points may raise precision quickly
[19:54:53] <alex_joni> psha: or degrade it
[19:55:04] <cradek> overdetermined != precise
[19:55:15] <alex_joni> 3 points = precise definition of a circle
[19:55:24] <alex_joni> more than 3 means you'll have to approximate
[19:55:38] <psha> alex_joni: it's simple to filter out 'outstanding' points
[19:55:49] <cradek> ??
[19:55:55] <cradek> you will guess which ones are wrong?
[19:56:29] <alex_joni> as a side note: on the robots I program, you always program a circle from 3 points
[19:56:30] <psha> if we beleive that user is giving us points with random offset then more points gives us ability to better approximate center
[19:56:30] <psha> 3 points is not precise since user is not giving us correct points :)
[19:56:31] <cradek> if you have 4 points you have 3 circles defined
[19:56:33] <alex_joni> and only 3 points
[19:57:03] <alex_joni> psha: say I click 4 points on the corners of a square
[19:57:06] <andypugh> Incomplete circle bores (motor mounts with belt clearance). I centred on the castings with a DTI (the other hole was not a good reference): http://picasaweb.google.com/bodgesoc/Gibbs#5511282568768390018
[19:57:08] <alex_joni> what is the "correct" circle?
[19:57:11] <psha> cradek: for 4 points it's not trivial but for 5 it is possible to filter outstanding point
[19:57:15] <psha> for 4 it's impossible :)
[19:57:46] <cradek> andypugh: I'd have turned it 45 degrees and used my method
[19:57:54] <alex_joni> hmm.. actually a square can be put inside a circle, so my example is useless :D
[19:58:12] <andypugh> With 4 points you do a least-squares estimate and return diameter and standard deviation.
[19:59:16] <andypugh> cradek: It would have to be exactly 45 degress to get the other bore in the right place.
[19:59:41] <andypugh> Though I was more saying that incomplete bores do occasionally happen.
[19:59:49] <cradek> I'm not sure what I'm looking at, but I see full circles under the incomplete ones
[20:00:05] <cradek> I didn't say they didn't - I said I had not run into that situation
[20:00:33] <andypugh> Raw castings, and I had barely enough allowance, the other holes were not precisely in the middle of the bigger ones.
[20:01:20] <andypugh> Also, you would smash the camera on the big holes trying to line up the little ones :-)
[20:02:24] <andypugh> But, the real point is that parlour-tricks are cool.
[20:03:08] <cradek> don't let me stop you guys - I just think sometimes software gets worse as you add more features, and the availability of the feature implies it's a good method to use.
[20:03:40] <psha> cradek: i'm just collecting ideas :)
[20:03:52] <andypugh> Which brings me neatly onto my current issue. 8i20 nvram programming...
[20:04:06] <andypugh> Make it work, or delete it completely?
[20:04:26] <alex_joni> andypugh: do the muntzing thing
[20:04:46] <alex_joni> http://en.wikipedia.org/wiki/Muntzing
[20:04:47] <andypugh> Crank handles on machine tools?
[20:06:17] <andypugh> Thing is, that currently the driver has one user, and he wants to change the current limit as he is using 2kW motors. So so far the evidence is that there is 100% demand.
[20:06:26] <cradek> haha
[20:07:10] <psha> heh, cam util has ~2 users :)
[20:07:27] <psha> and some 'potential' users :)
[20:07:51] <andypugh> Does Chris Morley ever hang around here? I want to chat about the idea of putting an interface to the onboard parameters into pncconf.
[20:07:54] <psha> so andy's request will satisfy 33% of users :)
[20:08:13] <psha> andypugh: i've tried to invite him here but failed :(
[20:08:43] <psha> it's sad since our communcation is limited to 2 mails per day - one from me and one from him :)
[20:09:10] <cradek> psha: 99% of what I do with my probe is finding an edge, finding the top surface, finding the center of a hole, finding the diameter of a hole
[20:09:29] <andypugh> I think I will abandon the setup mode, and put the driver in setup mode any time it is stopped (which is a pre-requisite to programming them)
[20:09:37] <cradek> psha: well 90. another common operation is measuring the left and right (or front and back) of something so I can find the center
[20:10:04] <andypugh> Do the cameras report back their focal length?
[20:10:19] <psha> current cam state is that there is adjustable circle to find center/diameter of a hole
[20:11:04] <psha> also it's possible to replace circle with square or two parallel lines for finding center/width
[20:11:06] <cradek> here is the left edge - here is the right edge - tell me the length and the coordinate of the center
[20:12:02] <psha> one very intersting application of camera on the mill is laser scanning :)
[20:12:04] <cradek> with the probe I just set origin at left side, measure right side, divide number in half and jog there, set origin again
[20:12:52] <cradek> unfortunately you can't measure depth/height/top surface origin. that's a very useful feature of a probe
[20:13:22] <psha> cradek: yes, that's not possible without additional stuff
[20:13:42] <cradek> you can line up to a scribe/punch mark easily though - can't do that with a probe - have to use a wiggler
[20:14:12] <cradek> although I really never do that - that's for manual machinists
[20:14:36] <alex_joni> meh, this is annoying: http://www.theregister.co.uk/2010/11/30/ransomware_trojan_returns/
[20:16:21] <cradek> meh, who cares
[20:17:04] <alex_joni> if you do sysadmin work, you start to care :D
[20:17:28] <alex_joni> (only with insufficient backup schemes though :)
[20:17:51] <cradek> if you do sysadmin work and have insufficient backups... gee ...
[20:17:58] <alex_joni> :P
[20:18:09] <cradek> then you used to do sysadmin work? haha
[20:19:39] <UncleG> If I have M64P1 > dwell .5 seconds > M65P1 how would I bundle that code into a single command like Mxxx ?
[20:19:58] <UncleG> Link to man page?
[20:20:06] <cradek> you can't
[20:20:29] <cradek> m64p1 / g4p.5 / m65p1
[20:20:49] <psha> o-call?
[20:20:52] <UncleG> bummer. maybe I could just write a call on line etc, command and put all those in a canned cycle and just call the canned cycle
[20:21:08] <UncleG> and just put them in every one of my files
[20:21:32] <UncleG> Its an air cylinder that flips my turret on my index machine
[20:22:08] <UncleG> I thought maybe I could provoke the turret flip with a single short command
[20:22:43] <cradek> M1xx could do halcmd sets whatever true; sleep .5; halcmd sets whatever false
[20:22:56] <cradek> you'd have to unhook it from motion.digital-out-whatever
[20:23:02] <UncleG> that would be fine
[20:23:09] <psha> UncleG: 'O<turretflip> call' is not one command? :)
[20:23:52] <UncleG> i have to do m64p1 / g4p.5 / m65p1
[20:24:26] <UncleG> The turret flip is activated by an air cylinder that extend, and engaged a trip dog
[20:24:41] <UncleG> engages*
[20:25:55] <UncleG> I think that should work for me Cradek~ can I link the action to a remote pendant for jog mode as well?
[20:26:46] <cradek> yes you could fire off a M100 with a halui mdi command
[20:27:50] <UncleG> Awesome, I think I can figure most of the rest of this out
[20:30:30] <andypugh> You could fire off an O<turretflip> call as an MDI command too, if you prefer.
[20:30:38] <cradek> not in 2.4
[20:31:11] <andypugh> You sure?
[20:31:20] <cradek> 90%
[20:31:44] <andypugh> Yeah, I find I am unsure what version I am running on the lathe/mill
[20:31:57] <andypugh> I think it probably is 2.5 and I have forgotten.
[20:32:51] <andypugh> Must, be, I went to 2.5 to develop the three-phase PWM stuff.
[20:33:03] <andypugh> Ignore me chaps! (Oh, you were?)
[20:34:43] <andypugh> Can you dynamically change the colour of a gladevcp bar? (I am pretty sure you can't with a PyVCP?)
[20:37:47] <psha> andypugh: whole color?
[20:37:47] <psha> via hal - no :) how do you imagine passing color over hal pin? :)
[20:37:47] <psha> with some custom hook - trivial, just set bg_color/z0_color property to what you want
[20:38:20] <alex_joni> psha: 3 pins for rgb ? :P
[20:38:27] <alex_joni> (just kidding..)
[20:38:50] <psha> alex_joni: 3 for bg, 3 for first zone... :) i've already added controls for crosshair color in camera app :)
[20:39:00] <andypugh> I wasn't asking for me :-). it was a forum query about pyvcp, and I was fairly sure the answer there was "no" but had no idea about the new, shiny toy.
[20:39:04] <alex_joni> 3 for color, 2-3 for muxing :D
[20:39:27] <alex_joni> select bg, fg, etc
[20:40:30] <psha> andypugh: it's possible to add some custom hooks but by default it's not possible...
[20:40:50] <andypugh> Thanks
[20:41:34] <psha> but when glib Pin wrapper will be ready it will be pretty easy to do
[20:42:42] <psha> and with mhaberler's userfuncs feature
[20:43:00] <psha> add custom handler and flip colors of bar to what you want
[20:43:25] <mhaberler> and it will wash the dishes!
[20:44:16] <psha> mhaberler: i'm not kidding :) if wisely used it's possible to do whole control gui with it :)
[20:44:46] <tlhiv_laptop> hi folks ... can someone help me get EMC running on a Gentoo system?
[20:45:08] <tlhiv_laptop> i've used Gentoo for years, and i don't mind compiling the kernel, patching the kernel, compiling software, etc.
[20:45:46] <tlhiv_laptop> it seems that EMC needs RTAI, but the latest version of RTAI seems only to suppor the 2.6.23 kernel ... that's fairly old now
[20:45:57] <psha> it's not correct
[20:46:23] <psha> EMC2 livecd is baed on 2.6.32 + rtai
[20:46:29] <tlhiv_laptop> ;) ok
[20:46:43] <psha> based
[20:47:03] <psha> i think you'd better stick with it
[20:47:10] <tlhiv_laptop> hmmm ... well the HAL patches that come with RTAI seem only to apply to (at most) 2.6.23
[20:47:35] <tlhiv_laptop> i'm talking about RTAI version 3.8.1
[20:48:35] <cradek> I think you pretty much have to get rtai from their vc to get a recent version
[20:48:48] <cradek> it's a big pain finding kernel and rtai versions that work correctly together
[20:48:51] <psha> i'm not confident in what's there on officail rtai page but have no problems in building 2.6.32 debian kernel + rtai-source
[20:49:34] <SWPadnos> http://wiki.linuxcnc.org/emcinfo.pl?Installing_EMC2#Preparing_other_versions_of_Linux_to_compile_emc2
[20:49:35] <tlhiv_laptop> psha: what version is rtai-source?
[20:49:36] <cradek> are you going to control a machine with this, or just simulate?
[20:49:42] <SWPadnos> may be out of date ...
[20:50:01] <SWPadnos> someone was working on an ebuild a couple of years ago, but I don't know what ever happened to that
[20:50:35] <tlhiv_laptop> cradek: to tell you the truth, i really don't know ... i am about to build my first CNC machine, and i'm still learning the fundamentals
[20:51:07] <cradek> tlhiv_laptop: you can experiment with emc in simulator mode without special kernel/rtai setup
[20:51:31] <cradek> if you later decide you want to control a machine, I strongly suggest using our kernel/distribution to avoid much pain
[20:52:00] <tlhiv_laptop> cradek: i don't mind pain ... in fact, i would really like (as a challenge) to get this working with Gentoo
[20:53:08] <cradek> ok I understand
[20:53:31] <tlhiv_laptop> i think i might be missing something ... my understanding is that i should be build my kernel with RTAI support (not sure exactly what that means really) and then build/install rtai
[20:53:36] <tlhiv_laptop> then i should be able to build emc
[20:53:54] <psha> and then you should descide where to put modules
[20:54:03] <psha> and where not to put them
[20:54:26] <psha> i've tried to build pure debian packages (not ubuntu ones) with FHS in mind and later agreed with cradek )
[20:54:32] <SWPadnos> IIRC, you also need to be careful about how you build EMC - the normal build process wants you to be running the RT kernel you're linking against
[20:54:36] <psha> it's not challenge but pain :)
[20:54:45] <SWPadnos> it's a little like the multi-stage gcc/libc build process
[20:55:03] <psha> SWPadnos: only for tests?
[20:55:33] <psha> SWPadnos: for build you only need installed haders
[20:55:34] <psha> headers
[20:55:51] <tlhiv_laptop> so just to be clear, if i do decide to do the Ubuntu install, does the LiveCD allow me to physically install to the hard drive, or will i be running everything off of this cd?
[20:55:53] <SWPadnos> no. I think it had to do with module search paths or something. at least in the "automatic, don't specify where the RT modules are" mode
[20:56:08] <cradek> it's a live + install cd
[20:56:16] <SWPadnos> tlhiv_laptop, either, you can test from the CD run and install if you like
[20:57:05] <psha> tlhiv_laptop: also don't forget that 'livecd' is not always CD :)
[20:57:15] <tlhiv_laptop> psha: right ;)
[20:57:16] <psha> you may write it to flash drive and be happy without installations :)
[20:58:10] <tlhiv_laptop> so my understanding is that i have to (1) build a kernel with RTAI support, (2) build/install RTAI, (3) build/install EMC
[20:58:17] <tlhiv_laptop> isn't this all that is needed?
[20:58:29] <psha> this is all
[20:58:42] <psha> (4) ensure that everything is working as desired
[20:58:47] <psha> that's most important part :)
[20:59:08] <tlhiv_laptop> so i guess i'm stuck in step (1) because i don't seem to have the correct "patch(es)" to patch my 2.6.35 kernel
[20:59:20] <cradek> (5) put up with us forever wondering if it's your build that's bad if you report bugs
[20:59:34] <tlhiv_laptop> hehe
[20:59:39] <cradek> tlhiv_laptop: explore the rtai website and mailing list
[20:59:46] <psha> for 32 kernel you may take ubuntu/debian patch
[20:59:52] <tlhiv_laptop> yeah i've been looking all day ;)
[20:59:59] <cradek> tlhiv_laptop: this is only tangentially the right place to ask for help building rtai
[21:00:02] <tlhiv_laptop> i am using 32 bit kernel
[21:00:31] <tlhiv_laptop> psha: where do i find this patch?
[21:00:45] <psha> http://packages.debian.org/rtai-source
[21:00:50] <psha> or from ubuntu
[21:01:17] <tlhiv_laptop> lenny, squeeze, or sid?
[21:01:39] <psha> squeeze/sid
[21:01:40] <cradek> try the third link on http://www.google.com/search?q=rtai+2.6.32+patch
[21:02:27] <tlhiv_laptop> hmmm ... so it sounds like i need to at least downgrade to 2.6.32 ;)
[21:02:53] <psha> tlhiv_laptop: if you don't link last ralink wifi or radeon drm it's ok :)
[21:03:45] <psha> s/link/need/
[21:03:48] <psha> strange typo...
[21:03:53] <psha> i'll better go to bed
[21:04:53] <tlhiv_laptop> seems that 2.6.32-r20 (gentoo) kernel is unmasked :-)
[21:05:55] <tlhiv_laptop> so let me get this straight ... this simple little few line patch is all that's needed to "prepare" the kernel for rtai?
[21:07:26] <SWPadnos> RTAI consists of two major components: an ADEOS-enabled kernel, and a set of kernel modules
[21:07:40] <SWPadnos> the little patch makes the kernel ADEOS-enabled
[21:07:56] <SWPadnos> the RTAI modules give you all the features of RTAI
[21:08:12] <tlhiv_laptop> and building/installing RTAI provides the modules?
[21:08:29] <SWPadnos> makes them available for when you want to run an RT application, yes
[21:08:48] <SWPadnos> the system is run in non-RT mode by default
[21:09:14] <SWPadnos> when you load EMC, various RTAI modules are loaded to provides the RT services that EMC needs
[21:09:24] <SWPadnos> those modules are unloaded when EMC stops
[21:09:32] <tlhiv_laptop> well this machine will be used ONLY for EMC
[21:09:44] <SWPadnos> you can probably configure it differently if you like
[21:09:56] <tlhiv_laptop> it's a dedicated Pentium 4 machine that we intend to control our CNC
[21:10:09] <SWPadnos> that's how the run script operates now, but you can change it if you like
[21:10:57] <tlhiv_laptop> on another (but not unrelated) topic, is there an IRC channel for general CNC talk?
[21:11:26] <jthornton> the forum has some general CNC topics
[21:14:51] <JT-Hardinge> my favorite is practical machinist my least favorite is the zone
[21:15:19] <SWPadnos> there's some general CNC talk here (and other unrelated talk even :) )
[21:15:26] <SWPadnos> I don't know if there's a #CNC or something
[21:16:12] <tlhiv_laptop> cradek: i'm not sure if that is the correct patch
[21:16:24] <tlhiv_laptop> cradek: that one seems like it's patching RTAI and not the kernel
[21:24:22] <cradek> if you're ok with emc-specific answers here, I think general cnc questions are fine
[21:32:13] <tlhiv_laptop> cradek: are these the patches that i need for the kernel? --> http://download.gna.org/adeos/patches/v2.6/x86/older/
[21:36:20] <SWPadnos> or this, if you want to keep 2.6.35: http://download.gna.org/adeos/patches/v2.6/x86/adeos-ipipe-2.6.35.7-x86-2.8-00.patch
[21:37:01] <tlhiv_laptop> SWPadnos: i would love to, but it seems that i can't build RTAI with this kernel version :-/
[21:37:10] <SWPadnos> oh. ok then :)
[21:37:25] <SWPadnos> I haven't bothered building my own kernels for quite some time
[21:37:39] <tlhiv_laptop> i'm just going to go with a vanilla version of 2.6.32.20
[21:37:41] <SWPadnos> (and haven't used Gentoo since the 2.4-2.5 changeover era)
[21:37:53] <tlhiv_laptop> and the corresponding adeos-ipipe patch
[21:41:12] <alex_joni> using adeos patches for RTAI only works sometimes
[21:41:49] <alex_joni> best is to use the patch from RTAI directly
[21:41:54] <cradek> RTAI only works sometimes
[21:42:10] <cradek> sorry - not helping
[21:42:13] <tlhiv_laptop> which is the more reliable?
[21:42:25] <alex_joni> for RTAI? use the RTAI patches
[21:42:44] <tlhiv_laptop> alex_joni: what's the more current RTAI patch?
[21:44:09] <alex_joni> https://gna.org/cvs/?group=rtai
[21:44:30] <tlhiv_laptop> seems down
[21:45:50] <alex_joni> yup
[21:45:59] <alex_joni> I think 2.6.35 should be stable-ish
[21:46:12] <alex_joni> there is a .36 patch, but no idea how it works
[21:46:12] <tlhiv_laptop> i can't find the correct patch for 2.6.35 though
[21:46:31] <alex_joni> should be in CVS under magma
[21:46:45] <alex_joni> (when it's working again, or you can checkout the source)
[21:46:52] <tlhiv_laptop> not sure how to "check out" the source
[21:47:15] <kb18951452> Anyone seen/used/heard of a DIT Pendant?
[21:47:21] <kb18951452> DIY*
[21:48:40] <andypugh> Yes, many of them.
[21:48:52] <andypugh> Simplest is to use a USB gamepad.
[21:49:57] <andypugh> Much less simple is http://ckcnc.wordpress.com/2010/11/20/arduino-emc-integration-how-it-works/
[21:50:48] <andypugh> How to configure a gampad is here.. http://wiki.linuxcnc.org/emcinfo.pl?Simple_Remote_Pendant
[21:51:08] <alex_joni> tlhiv_laptop: cvs co ...
[21:51:41] <tlhiv_laptop> alex_joni: building cvs now ;) ... freshly built gentoo system :-)
[21:52:44] <alex_joni> to each his own
[21:52:46] <alex_joni> :/
[21:53:24] <tlhiv_laptop> alex_joni: so am i right in assuming that the RTAI patch simply patches HAL?
[21:53:41] <alex_joni> depends on which HAL..
[21:53:58] <alex_joni> but not really, the RTAI patch is basicly an ADEOS patch
[21:54:04] <Valen> got mah cheap Chinese spinde on its way
[21:54:27] <tlhiv_laptop> alex_joni: but you said the adeos patch is not as reliable
[21:56:29] <andypugh> Valen: Eeek! I do hope you didn't spend money purely on my uninformed advice?
[21:57:02] <Valen> been tossing up for a while
[21:57:52] <andypugh> As I see it the bearings might be a weak spot, but they are the easiest parts to source.
[21:58:07] <Valen> low speed operation is also going to be an issue I reckon
[21:58:48] <Valen> cutting steel
[21:58:58] <Valen> unless we can get ceramic cutters
[21:59:08] <andypugh> I wonder of the included(?) inverter is a flux-vectoring type?
[21:59:19] <Valen> probably not
[22:16:34] <tlhiv_laptop> alex_joni: clearly the CVS has MUCH more updated patches than the "official" versions ;)
[22:45:10] <mshaver> andypugh: Regarding NVRAM on the 8i20 - I don't expressly _need_ to change these settings from within EMC. If I could, that would be great, but as long as I can do it some sort of way (even by booting DOS), and as long as there are enough pins & params to control the motor during machine operation, then all is well.
[22:45:40] <andypugh> Can you do it booting DOS?
[22:46:17] <mshaver> I think that's how Peter does it to begin with. _He_ must be able to do it somehow!
[22:46:43] <andypugh> I think he does it with very fine tweezers directly in the EEPROM
[22:47:10] <mshaver> I'm assuming that none of these NVRAM settings are required to change "on the fly".
[22:47:20] <mshaver> hah!
[22:48:10] <mshaver> The next point where I'll likely need your help is with Hall switches.
[22:48:32] <andypugh> There are a limited number of writes allowed to the actual EEPROM
[22:49:12] <mshaver> If I can stop goofing off, I'll get then hooked up in a couple days& then we'll see if homing is made more reliable.
[22:50:35] <andypugh> But I am looking to make it possible to alter some other things semi on-the-fly (you need to stop the sserial port to programme the card)
[22:50:36] <mshaver> True, but it's in the 10s or 100s of thousands. I think you only need to set those once - either the same for each machine/motor, or on a per machine/motor basis (per model or per unit).
[22:50:58] <mshaver> That was remarkable ambiguous...
[22:51:14] <andypugh> I am rather thinking that pncconf is the natural tool for actual nvram settings changes
[22:51:42] <mshaver> what's that? wait, I'll look it up...
[22:52:03] <andypugh> And pncconf can actually do all the setup in RAW mode.
[22:55:25] <JT-Hardinge> is there a way to specify the size of the Axis screen at start up?
[22:56:35] <JT-Hardinge> wow, that part took 20 minutes and 3 seconds to make...
[22:56:59] <mshaver> Wow (looking at http://wiki.linuxcnc.org/emcinfo.pl?Pncconf) I'm not really a gui configurator user, but that looks really nice! If that where the NVRAM setup should go, then that's fine with me!
[22:57:03] <Valen> andypugh: so you use a chinese spindle?
[22:57:24] <mshaver> JT-Hardinge: devilspie
[22:58:14] <Valen> or anybody else for that matter
[22:58:21] <andypugh> Valen: I guess so, but not in the sense you mean. I have a nasty chinese mill, with a nasty spindle. (plastic gears, DC motor with exploding commutator)
[22:58:50] <Valen> just ordered one of these http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=250532286917&ssPageName=ADME:X:RTQ:US:1123
[22:59:01] <Valen> it'll go with the cheap chinese everything else I spose
[22:59:59] <andypugh> I also have a nice English spindle, with bearings hand-engraved by Frenchmen. Mainly because the other one turned out so nasty.
[23:05:25] <andypugh> Valen: That does seem very fast for steel.
[23:05:46] <andypugh> Is steel the main use, or an occasional thing?
[23:08:18] <JT-Hardinge> mshaver: does devilspie interfere with EMC in any way?
[23:09:13] <mshaver> not that I have seen - I've recommended it as a way to maximize AXIS at start up
[23:09:44] <andypugh> Valen: Actually, using numbers for carbide cutters it's about right.
[23:10:58] <andypugh> 1/2" tool, 900FPM (Top end, but lets assume lots of coolant) is 6875rpm, and for a 1/4" cutter you are right where you want to be.
[23:11:51] <mshaver> devilspie info: http://www.pastebin.ca/2007159
[23:13:56] <JT-Hardinge> mshaver: thanks
[23:15:06] <Valen> 6000rpm is its absolute minimum
[23:15:22] <Valen> and we want to do stainless and titanium as well
[23:15:34] <Valen> so it'll be getting warm at 6000RPM probably
[23:15:37] <Jymmm> I feel sorry for the guy that had to climb the palm trees http://ireport.cnn.com/docs/DOC-523097?hpt=Sbin
[23:16:20] <Valen> I wish there was a way of telling it not to download videos
[23:16:28] <Valen> let me read it, then I'll decide
[23:17:10] <andypugh> Web site is broken, there is a window over the video pane, but the "close" button for that window is hidden behind a Subaru advert.
[23:22:18] <kb18951452> Anyone see/used a DIY Pulse Generator?
[23:28:16] <andypugh> I have made my own encoders, if that is close enough.
[23:35:23] <JT-Hardinge> I made one in HAL once
[23:39:04] <JT-Hardinge> is there any instructions for installing EMC sim in 10.10 anywhere on the net?
[23:41:57] <Jymmm> isn't it already there?
[23:42:07] <Jymmm> or is that 10.8
[23:44:14] <JT-Hardinge> I found 10.04 but not 10.10
[23:46:44] <tlhiv_laptop> what graphical libs does emc build against? GTK? QT? other?
[23:49:29] <JT-Hardinge> I assume it depends on what GUI you choose
[23:53:59] <Jymmm> JT-Hardinge: since it's just sim, can't you instal it like anythign else?
[23:54:12] <Jymmm> even LIVECD has sim iirc
[23:55:43] <JT-Hardinge> I don't know, just trying to help greybeard out on the forum
[23:56:08] <Jymmm> JT-Hardinge: I'd wait for alex to show up later, I think he would know
[23:56:52] <Jymmm> Ok, if 16 sq in of pellets stop the air, you think 48 sq in of surface area would be enough?
[23:58:16] <JT-Hardinge> for what?
[23:59:34] <Jymmm> JT-Hardinge: In my air scrubber, I have two cardbaord boxes stacked on top of each other filled with 4mm carbon pellets. the intake is 4" duct. The airflow moves into bottom box, thru the top box then out. But,.....