Advanced Search

That 'relkins.c' is a bit more effort than I'm willing to put into it.

Maybe have a look at the rather extensive comments in the source code for matrixkins, there are examples no how to set the parameters:
github.com/LinuxCNC/linuxcnc/blob/master...ents/matrixkins.comp

That is the first ever GNOME GUI on some GNU Linux version on that document.
Also searching that document for windows or NT shows nothing.

Strange
Checking the halui.mode inputs and the settings from the custom_postgui.HAL are showing up but the output states are not matching
Even when setting the halui inputs manually (using the halcmd) to "re-trip" them, the output states (is-teleop and is-joint) don't change looking either at the command line or in the HALshow within the program.
Setting those input values doesn't actually trigger the output change, something else must be happening when an MDI command happens that doesn't happen when I set the mode values.

More digging

I don't know where I heard about Windows NT, there isn't any mention of it here:
tsapps.nist.gov/publication/get_pdf.cfm?pub_id=821651
However the very first implementation described there seems to have run the GUI on a Windows PC.

You can create an issue here: github.com/jethornton/mesact/issues

Parallel ports share much in common with GPIO pins so I wasn't sure. Thanks for the response and particularly the references.

Look, I think you're missing my point.

I'm just saying that if we can get GitHub to build and provide downloads links and a download page for release DEBs, that's good.

Our DreamHost server will save some bandwidth (as GitHub provides the DEB download), which should speed up the download of ISOs (a little bit).

I think that should work with a couple conditions:

1. You must have a digital spindle enable signal because the spindle output will default to -5V
when LinuxCNC is not running or after a watchdog fault

2. You need to set the PWMgens offset mode "true" so that a 0 PWM value results in
a 50% duty cycle and therefore a 0V output
 

That's what I figured but I can't seem to get it to work.
I've tried
setp halui.mode.in-teleop 1 but that's an output so it errors as not writable
setp halui.mode.is-manual 1 but that's also an output so it errors as not writable

Tried
setp halui.mode.manual 1 doesn't seem to do anything
setp halui.mode.teleop 1 doesn't seem to do anything
setp halui.mode.enabled 1 No such pin error

Looking at the states when the program loads and I can't use the MPG
10  bit   IN          FALSE  halui.mode.auto
    10  bit   OUT         FALSE  halui.mode.is-auto
    10  bit   OUT          TRUE  halui.mode.is-joint
    10  bit   OUT          TRUE  halui.mode.is-manual
    10  bit   OUT         FALSE  halui.mode.is-mdi ==> MDI-mode
    10  bit   OUT         FALSE  halui.mode.is-teleop
    10  bit   IN          FALSE  halui.mode.joint
    10  bit   IN          FALSE  halui.mode.manual
    10  bit   IN          FALSE  halui.mode.mdi
    10  bit   IN           TRUE  halui.mode.teleop
 
and after I make some kind of MDI move

10  bit   IN          FALSE  halui.mode.auto
    10  bit   OUT         FALSE  halui.mode.is-auto
    10  bit   OUT         FALSE  halui.mode.is-joint
    10  bit   OUT          TRUE  halui.mode.is-manual
    10  bit   OUT         FALSE  halui.mode.is-mdi ==> MDI-mode
    10  bit   OUT          TRUE  halui.mode.is-teleop
    10  bit   IN          FALSE  halui.mode.joint
    10  bit   IN           TRUE  halui.mode.manual
    10  bit   IN          FALSE  halui.mode.mdi
    10  bit   IN           TRUE  halui.mode.teleop


I see that the halui.mode.is-joint went from true to false. And probable a coincidence but the last thing I did was to get the MPG to work AFTER a MDI move I also changed the references for the MPG from joints(from a previous setup I did years ago)  to axis references.
Guess I'll try setting that to false as well and see what happens

Thanks for the help.
M
 

Thank you, however, this is not necessary. If you already have a Matrixkins component built into the LinuxCNC kernel, then you need to use it. According to the description, it is quite suitable. The main thing is to understand the configuration synatxis for hal.
According to the description, the component built into the kernel solves issues of three axes at once. Simple (linear), but closes.

I don’t have the realkins component itself. The forum where this component was published has already died. The module is here. That's exactly what's holding me up on version 2.8.4

If you attach the code of your current 'relkins' component I can have a look at updating it to be used in version 2.9

TWOPASS was already on.

From the description at the link I see the following alorhythm -

1. I know the amount of deviation
2. in the case of XU, I need to explicitly specify the parameter for the formula -
X_joint = C_xx * X_axis + C_xy * Y_axis + C_xz * Z_axis

Those. I know the coefficient from clause 1
C_xx = 1
C_xy = (deviation X divided by width Y)

But what are the rules of use, syntax, examples?

loadrt matrixkins count=0
C_xx = 1
C_xy = 0.03
or
C_xy = [3/100]

or otherwise?
how to change the C_xy parameter on the fly without rebooting the program?
if loading a component requires specifying the contact number count=N, then N is what, what are the options? And where should this count be specified in the C_xy job?

In general, we need examples of syntax and application
In man, the example section of the configuration file shines with emptiness

I just worked through the Ben Fleming circuit (not 100 % yet). It seems that the voltage is ~ 100 V DC and the current can be limited by power resistors. A couple of mosfets get pulsed by a 555 timer, the pulse duration is set via pots. I replicated part of the circuit in LTSpice an ran a couple of different pot settings to get an idea of the pulse on and off duration. Also compared these values to commercial machines. The calculated frequency is ranging from 0,5 kHz to 60 kHz. Do I even need the 555 timer and the two pots or could I use PWM signal from linux to do that? A mesa card can go up to 386 kHz so should be plenty of room for that.

tommylight wrote:

All those can be easily done with any microcontroller, but might need better ADC converters.

That would be important for the voltage gap feedback. Would an off the shelf voltage to frequency converter work for that?

Hola a todos. Tengo una CNC3040 con motores nema 23 con electrónica mesa 7I96 y husillos de bolas recirculantes.
    la utilizo para la fabricación de prototipos de placas de circuito impreso o PCB, pues antes de empezar ha hacer el trabajo tengo que subir y bajar muchas veces el eje Z para que corte a la medida exacta... Por ejemplo. 
Mando el comando para que los ejes se posicionen en el origen de la pieza y le cargo el archivo gcode y le doy a empezar. Pues observo que no está cortando a la altura correcta, vuelvo a mandar los ejes al origen máquina y vuelvo a darle a comenzar el trabajo y vuelve a cortar a la altura incorrecta y vuelvo a repetir el ciclo hasta que en una de esas ya corta a la altura correcta y ya no vuelve a perder la altura y termina el trabajo perfectamente, en los ejes X e Y no le pasa eso ¿Alguien sabe que le pasa al eje Z y como podría corregir el problema?.
Saludos