Advanced Search

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

sounds really good, thanks for the answer.
Are there any configuration examples somewhere? that loadrt is clear.

The current version 2.8.4 I can -
1. enable
2. launch a geometry check program, which, based on the results of the Sonde of the first class square, will calculate and tell me the amount of deviation from the perpendicular
3. Without turning off the program, I can enter a command like - in the terminal -
halcmd setp relkins.adjX.Y1030 -1.82

and get the corrected geometry without restarting the program.


How to do this in version 2.9? are there any examples?

I'm not familiar with 'relkins' but it will likely need updating due to the addition of the switchable kinematic feature. Discussion:
forum.linuxcnc.org/49-basic-configuratio...-xyz?start=10#311322

However there is a hal component included in version 2.9 that should handle all kinds of skew corrections:
www.linuxcnc.org/docs/devel/html/man/man9/matrixkins.9.html

By default it is not possible to 'loadrt' the same hal component twice.
Try enabling TWOPASS:
linuxcnc.org/docs/html/hal/twopass.html

I try to integrate my gamepad (which is working under axis) into Probe Basic.
When I read the error report correct "not: already exists" is the complaint.
I tried to format the not function in an other way in the probe_basic_postgui.hal, but that doesn't work.
Even if I only add a simple halfile = not.hal (instead of gamepad-encoder-16.hal) with a single not function I get an error.

  Does this diagram apply to a bipolar connection?
www.aliexpress.com/p/tesla-landing/index...aXGaNZQaAvhREALw_wcB

I have now installed version 2.1 which accepts the official version of mesaflash 4.3.6... and I have no problem with exporting the configuration, but mesact messes up the configuration and here are a few things
1: does not export the given maximum velocity
2: export joint 0 instead of x for 7i77
3. will not enable drivers trough ena+ ena- on 7i77
4: instead of recognizing the pins it exports something like 7i96s.0.7i77.0.4 analogout instead of 7i96s.0.7i77.0.1
5.i/o also exports 7i96s.7i77.0.3 instead of 7i96s.0.7i77.0.0
and so on... I can't get a configuration that otherwise works with pnccfg normally, except that I would have to manually enter everything related to 7i96s, so in the end I will have to do that. I don't know how to post what you suggested on github so that the errors can be corrected. sure thank you very much

SOLVED  -
matrixkins is a standard technique for correcting the curvature of the portal in the LinuxCNC 2.10 kernel (and above). The issue of mutual correction of XYZ coordinates is resolved by specifying correction coefficients. However, this component is provided for single-engine machines, i.e. XYZ schemes. To use this software correction for LinuxCNC2.9 and possibly 2.8, you must manually compile and install the matrixkins module (create a text file matrixkins.comp and copy-paste text from github.com/LinuxCNC/linuxcnc/blob/master...ents/matrixkins.comp + sudo halcompile --install matrixkins.comp), and you must make changes to the ini machine configuration file. Hal file does not require correction. MAN does not give these examples. However, in the ini file 
instead of -
KINEMATICS = trivkins coordinates=XYZ
set -
KINEMATICS=matrixkins
then explicitly specify non-zero correction factors, for example -
halcmd setp matrixkins.C_yx -0.00252427184466  

You will find the coefficient values and the transformation matrix in MAN.
The halcmd setp matrixkins.C_yx correction command -0.00252427184466 can be executed in the terminal (not while the machine is moving!), i.e. the parameter can be changed on the fly.

In case you have a multi-engine XXYZ configuration, etc., you need to change the source code of the matrixkins component ( github.com/LinuxCNC/linuxcnc/blob/master...ents/matrixkins.comp ), similar to  www.forum.linuxcnc.org/38-general-linuxc...xyyz?start=20#338894

+ sudo halcompile --install matrixkins.comp

you need to be in teleop mode to jog so I think selecting MDI mode disable this. Check sate of halui.mode.is-teleop 
and select teleop mode by setting halui.mode.teleop to true

added them to the first post, pretty messy but that's where I'm at so far

Any issues relating to this please post here as they apart of the RPi Trixie build
github.com/ozzyrob/pi-gen-linuxcnc

Twice the money is still zero.

Your major issue is a lack of knowledge of how things work. First being not understanding how directory listings are produced by a web server. Secondly not understanding how a Debian repo works, and thirdly not bothering to find out how these things work.
Hoenstly, and I'm trying to be as diplomatic as I can, you have a low knowledge of how and why things work and yet this doesn't stop you giving out adice based on your knowledge. A classic example was when you suggested apt was stupid or frustrating or something similar for not installing packages due to version mismatches, this is to protect the system (and user) from a broken system.
Another thing that is an issue is when you make statements that you have no evidence for such as BSD has low latency, which also points to a lack of understanding of latency and jitter, hopefully you followed the link I provided to an explanation of latency within the context of the Linuxcnc ecosystem.
The thing that hurts the forum is that "member status" is based on the amount of posts. So when you give advice that is wrong or similar it is more hurtful than a recently joined member.
This is not a preseason issue, tho I have made some, I hope offensive remarks regarding yourself, but rather an issue that affects the quality of information that is given in the forum.
Yes I understand we are not born knowing everything, and the more you learn the less you tend to realise you know. Sometimes I am cagey about making sweeping claims, as I know Andy is aware of something that refutes what I said if it was made as a sweeping claim. And I'm thankful for that, I learn a little bit more. We all get corrected no one is special.
So what I'm trying to convey is if you sit back, observe, read and either make mental notes or notes in a notebook or similar and refer to them when answering you will become a member who's post have value. If not you will be seen as just a nuisance.