Advanced Search

Maybe this helps:
linuxcnc.org/docs/html/gcode/o-code.html...outine_return_values

Hello everyone
I am encountering an issue that appears to be EMI-related despite applying the usual mitigation steps.
EMI / System Setup:
VFD in steel enclosure
Input EMI filter
Ferrite rings on VFD output and all related cables
Shielded spindle/motor cables (grounded both ends)
AC/VFD cabling physically separated from DC lines
High-current DC cables are shielded and routed separately
Control Setup:
I am using the 7i96S encoder pins for MPG inputs (I have no motor or spindle encoders).
I do not have any unused OUT pins and only 4 available IN pins.
To get X, Y, Z, X1, X10, and X100 signals, I wired them to Raspberry Pi GPIO using external 10k pull-ups.
Issue
After wiring the MPG inputs, some of the pins intermittently remain latched HIGH or LOW, making jog selection unreliable. This problem has occurred twice, and both times it started after spindle-speed testing.
Earlier in setup, I believe I damaged one of the 7i96S step/dir channels. After that, instead of wiring STEP- and DIR- to GND and STEP+ / DIR+ to the respective inputs, I reversed them — STEP+ and DIR+ to +5V and STEP- / DIR- to the Mesa pins.I have since added additional ferrites and shielding, Ferrites on servo lines, data cables, and MPG cable
Steel mesh added to enclosure ventilation fans but the issue persists.
I am now planning to use the P1 pins on the 7i96S for X/Y/Z and X1/X10/X100 inputs instead.
W6 jumper is OFF.
Questions

1. The MPG has extra wires that appear intended to power LEDs or similar features. Can these be safely powered from the 5V rail, or is that risky for the inputs?
2. Where exactly are pin 1 and pin 2 physically located on the P1 expansion header?
3. Which P1 pins are appropriate to use for these MPG signals?
4. May I safely use 5V for the MPG input power in this configuration?
5. Any recommendations to avoid EMI issues when using these inputs on the 7i96S?
   
   Thank you for your help.
   
   
   

Hmm I wonder why apt purge worked in my situation ?

And why did it work when I created a new ISO and added a line to pressed that ran apt purge to remove the rasp firmware at the end of the installation process ?

here are  my current working files.

I have changed the loadrt line slightly  and receive a different error ? 


RUN_IN_PLACE=no
LINUXCNC_DIR=
LINUXCNC_BIN_DIR=/usr/bin
LINUXCNC_TCL_DIR=/usr/lib/tcltk/linuxcnc
LINUXCNC_SCRIPT_DIR=
LINUXCNC_RTLIB_DIR=/usr/lib/linuxcnc/modules
LINUXCNC_CONFIG_DIR=
LINUXCNC_LANG_DIR=/usr/lib/tcltk/linuxcnc/msgs
INIVAR=inivar
HALCMD=halcmd
LINUXCNC_EMCSH=/usr/bin/wish8.6
LINUXCNC - 2.9.6
Machine configuration directory is '/home/cnc/linuxcnc/configs/probe_basic_dev'
Machine configuration file is 'EL8_machine.ini'
INIFILE=/home/cnc/linuxcnc/configs/probe_basic_dev/EL8_machine.ini
VERSION=1.1
PARAMETER_FILE=linuxcnc.var
TPMOD=
HOMEMOD=el8_homecomp
TASK=milltask
HALUI=halui
DISPLAY=probe_basic
COORDINATES=X Y Z
KINEMATICS=trivkins coordinates=XYZ
Starting LinuxCNC...
Starting LinuxCNC server program: linuxcncsvr
Loading Real Time OS, RTAPI, and HAL_LIB modules
Starting LinuxCNC IO program: io
Starting HAL User Interface program: halui
linuxcnc TPMOD=tpmod HOMEMOD=el8_homecomp EMCMOT=motmod
Found file(REL): ./hallib/EL8_machine.hal
Shutting down and cleaning up LinuxCNC...
Removing HAL_LIB, RTAPI, and Real Time OS modules
Removing NML shared memory segments

Debug file information:
Note: Using POSIX realtime
HAL: ERROR: function 'cia402pv.Spindle.read-all' not found
./hallib/EL8_machine.hal:18: addf failed
2975
Stopping realtime threads
Unloading hal components
Note: Using POSIX realtime
 

I've been trying to develop a routine to repeatedly measure the four corners of my X/Y table and its center for leveling purposes.  The attached file is currently working great though it only does one set of measurements each run.  The M100 code at the end of the routine is a small python routine to read the proberesults.txt file, converting the values to floats and storing them into an excel spreadsheet.

Having had success with the routine I modified it to wrap the 5 measurement calls and the M100 call in a WHILE LOOP.  However, only the last set of measurements was posted to the spreadsheet. To get around this issue I would like to do the following:

1. Instead of writing each measurement to proberesults.txt, I would like the o200 subroutine to return the value of #<_z> for each measurement such that I can capture them and write a single line with all 5 measurements; i.e. (LOG,#<LF>,#<RF>,#<LR>,#<RR>,#<C>)

I capture #<_z> in a local variable, #<result> = #<_z>, then return it with:
o200 endsub [#<result>].

I've tried to assign the return value with:

#<LF> = o200 call p1,p2,p3,p4

and

o200 call p1,p2,p3,p4
#<LF> = #<_value>

and

#<LF> = [#<_value>]

All of these generate errors when I load the program file into LCNC.

What is the correct syntax for assigning return values from an oWord routine?


Kind regards,

Hakan wrote:

Chat gpt hehehe.
Go into halshow and list all spindle pins.

Read the motion(motmod) documentation, the spindle section linuxcnc.org/docs/html/man/man9/motion.9.html

Look at one of the examples in linuxcnc .
Pins created by mb2hal are just as good as any other pin.

in fairness to chat gpt it has helped me waste time.  

Halshow shows all the spindle.xxx pins but they are not writeable. I've looked at many examples, when I apply them to my config they don't work. I get pin doesn't exist on most things. That's why I am here asking questions. OBVIOUSLY things are missing from my main config, but I cant see what. I will collect my main hal and ini and mb2hal files that "work" in a moment and post them. It starts with no errors, and xyz move, and via halshow i can start the spindle, but nothing I do allows to to use m3 to start the spindle.

my1987toyota wrote:

If people wish to swallow the continued escalation of data mining the populus and participating
in police states. that is for them to decide.

  But to put it simply rodw you don't have to log into forums to READ the information. At least not yet.
I wont be surprised if that changes soon as well. You know for the kids. Wont somebody please think
of the Children. So why Should I have to log in to VIEW a video, that's my point.
I am tired of the increasing collection of data on everyone not just me.
I may not be able to stop the big machine, But I don't have to help it along either.

I cant find any reference to having to log on to view videos other than random people saying they have technical issues going back YEARS. Can you direct me to the youtube policy for this? I know this is is a new policy in the usa (maybe only in some states, not federal, and nothing to do with youtube of course) but i cant see where its stated by youtube as a rule. Got links?

I am currently watching videos while not logged in...

Thank You Very Much

Yes.

It been a long time since I have used classicladder but I think you need to link the set coil to a contact like in this example:

 

So you would connect your '%Q0' Output to the 'toggle' contact. (Like  'reset' in the example above but without the 'Time-off' block.)

Chat gpt hehehe.
Go into halshow and list all spindle pins.

Read the motion(motmod) documentation, the spindle section linuxcnc.org/docs/html/man/man9/motion.9.html

Look at one of the examples in linuxcnc .
Pins created by mb2hal are just as good as any other pin.

But there must be a reason that linuxcnc won't find it.
It looks okay in your files.
Response is exactly like it is commented out.
I tested to comment out the loadrt line and got exactly thisI also tested to mis-spell cia402pv.Spindl in the loadrt line and got the same error.
Also mis-spelling on the addf line give the same error.
There is some mis-spelling between the loadrt and the addf lines
but I don't see what it is.

From the output messages this is the hal file linuxcnc uses
./hallib/EL8_machine.hal
That's the one you showed, right?

I don't know. Can you retype those two lines?

hi there,

Just for anyone who searches and finds this, as of this writing (October 30 2025), if you have already run sudo apt update and sudo apt upgrade, you can't just run apt purge, you have to run:


And that will remove the error.

In addition, at least on my machine, after you run pncconf, you have to comment out lines 21 and 22 in order for LinuxCNC to launch.

Hi Peter,

Thank you for this information. Forgive me, I'm new. Are you associated with Mesa Electronics?

Regards,
John