Advanced Search

I'm not trying to dissuade you from using LCNC with an external PLC, but if you must have a physical PLC handling IO and similar...have you considered a PLC-only solution?

Some PLC's these days are available with high-speed outputs and built-in motion control (trajectory planner) functions.

I have built two automation machines recently using an Automation Direct ClickPlus PLC that drives three axes of coordinated motion via step/dir servos.  Programming was dirt-simple and the resulting motion met our expectations.

The main PLC has the HS outputs - no need for expensive daughter boards just to get the motion control features.

In my case the motion profiles are very simple; I've not explored the full capabilities of these PLCs... but if your needs are also fairly simple something like this might suffice.  Admittedly, the Click PLCs are fairly 'low-end', so higher-end PLC's likely have more sophisticated onboard motion control features.

For step/dir, Mesa 7i96S is the right choice, and if you do need more IO add Mesa 7i84 with 32 inputs and 16 outputs and wires to 7i96S with a simple cut down LAN cable.

TLDR
LinuxCNC has it's own PLS named ClasiLadder.
Either way you look at it, it is much easier to do everything in LinuxCNC as you can add VCP panels and tabs to Axis GUI to your hearts content.
Or am i missing something?

I've not noticed this behavior, but I'm not omniscient, I'd need more information, your pictures are lacking context...

Hello,
I'm new to Linuxcnc, I've been studying it for a month and I'm very happy with it. 
I'm retrofitting a Cortini mill (a very small industrial-grade italian cnc mill) and thanks to this forum I could get a Ethercat system working with Beckhoff hardware and I could add a lot of external hardware buttons to my system.
For example I can have a program run, pause physical button in auto mode.
I cannot have a physical button for the go/run mdi command line.
Please note: this has nothing to do with custom MDI macro buttons.
To be clear: I put the system in MDI mode, type  "G0X10" then I must use the mouse to have it run.
The keyboard enter button is not the same becouse if i recall a previous mdi command from the list touching it on a touchscreen (or selecting with the mouse) the Enter key doesn't work.
I could not find a halui connection to this go button.
The goal would be to connect it to the start program physical button that is already  working in auto mode but this could be a further step. It would just be great to connect it to a new physical button.
Is this in any way possible?
I hope to have been clear despite my bad english.
Many thanks
Alberto
 

@grandixximo 

Is it possible that if the maximum speed is not reached (without t3/s3), the jerking exceeds the limits?


   

gruß
Rüdiger

Here is a link to download everything I used to create the touch-off side panel. No installation needed as it just a python script and some support files. I am NOT a developer and this is totally ad hoc software. Also, some functionality is incomplete and if used you should consider it as a template to your specific needs. No instructions are included but the code is well commented and if you are even a novice in python you will have no issue editing it to your particular needs. This was tested on a pi5 8GB running Debian Trixie from the official downloads.

You will need Inkscape to Edit the GUI. I wrote a function in the script that opens and parses an Inkscape file and maps tkinter widgets to the named objects. The Inkscpape objects contain the X and Y position data and the width and height information for the tkinter widget. Doing this allows for super easy GUI creation and edits. The script comments explain this in more detail.

I am not offering support if you decide to use this, hopefully the community can make it better and share with everyone. I am SURE that my mdi commands and general code structure can and should be improved before considering this as a stable and good addition to LinuxCNC.

drive.google.com/drive/folders/1mLqdm_YS...5YEtc?usp=drive_link

To add some more technical context to our use case:

This is a large-format gantry-style machine for laser cleaning and automated spray painting of heavy equipment and large structures. The focus is on robustness, safety, and simple, repeatable operator workflows rather than traditional CNC part programming.

Our current concept is roughly:

• PLC: machine sequencing, state machine, safety logic, interlocks, recipe management, part dimensions, operator interface (HMI), and maintaining machine state independently of the PC
• LinuxCNC: motion control, path execution, coordinated axes, and low-level servo/stepper control

The PLC would command LinuxCNC at a higher level (start cycle, select program, pass parameters, etc.), while LinuxCNC would not be responsible for overall machine state or safety sequencing.

We are particularly interested in hearing from anyone who has implemented a similar architecture with a real industrial PLC + HMI, and how you structured communication, parameter passing, and responsibility boundaries between the PLC and LinuxCNC in production machines.

Any concrete examples or lessons learned from industrial environments would be very valuable.

Thanks for the detailed explanation — that is a very interesting architecture and I appreciate you taking the time to describe it.Thanks for the clarification regarding Beckhoff I/O. Just to be clear, in your setup you are using Beckhoff EtherCAT I/O as remote I/O, but without a Beckhoff PLC runtime (e.g. TwinCAT PLC), correct? In our case, we are specifically looking for architectures where a real PLC runtime (IEC 61131, ladder/ST, etc.) is responsible for machine sequencing and state, and LinuxCNC is used primarily for motion and interpolation.
In our case, one of the main drivers for using a PLC is to keep machine sequencing, state management, and operator workflow independent of the LinuxCNC PC. The goal is that the PLC maintains the machine state and basic operation even if the LinuxCNC PC or software needs to be restarted or serviced.

We are also targeting a more traditional industrial maintenance environment, where ladder/IEC PLC logic and a dedicated industrial HMI are preferred for long-term support by technicians who are not Linux or software specialists.

For that reason, we are specifically interested in experiences using a real PLC (e.g. Mitsubishi, Omron, Siemens, Keyence, Beckhoff, etc.) alongside LinuxCNC, and how people typically split responsibilities between PLC and LinuxCNC in that type of architecture.

Still, your approach with Node-RED, distributed I/O, and web-based HMI is useful as a reference for alternative architectures — thanks again for sharing your experience

I guess an important information is, i plan to run the drives in position control.
I plan to use the Yaskawa in closed loop.

So the drives would receive STEP and DIRECTION signal. According to manual i also has a completion signal called COIN. And obviously some alarm and enable signals as well.

Hope this clarifies a bit.

Dear LinuxCNC User,

I am using Mesa 7i95 board to run a single axis AC Servo motor. I want to control feed override externally using LabVIEW while the LinuxCNC G-code is running at programmed feed rate. This is my part of PhD research work to control the feed override of a running G-Code based on some unstable/chatter vibration in the process.

As I now aware that, there is no analog input pin available in Mesa 7i95 board such that I feed analog voltage and bind that pin with mpg feed override inbuilt pin of LinuxCNC software. From the forum, I got suggestions that is it possible to use encoder pwm pin for the same purpose. The idea is, generate a digital  signal A and B of 90 phase lag with LabVIEW and send it to encoder pin of 7i95 using LabVIEW cDAQ module NI-9401. So, I have generated two digital signal called A and B and create logic as following 
A-0, B-0;  A1, B0;  A1, B1; A0, B1. This logic clearly generated two A and B digital pwm type signal with 90 degree phase lag. I checked that in oscilloscope and it shows perfect squrae signal of 5 volt in high and 0 volt in low with 400 mV noise.

Now the following wiring I have done from NI-9401 to 7i95 encoder 2 slot.
A signal or digital output 1 of NI-9401 to A+ or pin 1 of 7i95
B signal or digital output 2 of NI-9401 to B+  or pin 4 of 7i95
GND of NI-9401 to GND or pin 3 of 7i95

After this connection, I checked the encoder count of encoder 2 in Hal show configuration. I observed that, the count is increasing, then decreaseing, and then again increasing, but overall it increases. I see it randomly increasing or decreasing. I played with loop timing of digital signal generator from 1 second to 1 mili-second but no improvement. I was expecting that if I set 1 sec loop time the counter of encoder will increase by 4 times because it is in quadrature mode. But again important point is I have not used the other encoder pin of encoder 2 slot. There are 8 pin available in encoder slot, A+, A-, GND, B+, B-, 5 Volt output, Z+, Z- respectively from pin 1 to pin 8. Out of the eight pin I am using only 3 pin.

Truly speaking, I don't understand how the mesa driver actually counts the quadrature pulses and how the motor encoder generate those pulses. At what frequency motor encoder generate those pulses and what frequency mesa drive read those pulses.

I believe either I am doing wrong wiring or setting wrong frequency to generate the signal.

I request expert to please guide me in doing proper wiring, frequency setting of loop in LabVIEW, HAL code setting in .hal and .ini file to get stable counter in encoder 2. After I get stable counter and full control on number of pulses, I will bind that pin with feed override pin and control the feed override externally using LabVIEW. From LabVIEW I will send fix number pulses either to increase or decrease the feed override. I will scale those pulses to match the feed override required value.

If this is possible let me know, else suggest me a different approach to do the same task either by buying new most suitable daughter board that supports with 7i95 and has analog input and output pin or any other way.

Thanks 
Suraj Kumar 

Newbie here! Be gentle

I have recently acquired a bunch of Omron Yaskawa servos and SGDH drives from a running machine.

Read the LinuxCNC documentation a while ago planning for a CNC setup for my minor lathe.
Searched the forum for experience with the Omron Yaskawa SGDH controllers. One guy doing mesa to control the SGDH.

First of all i would like to do x-y axis operations to do threads etc. Long term perspective i would like to do 3 axis machining, making the spindle A-axis servo controlled - and mounting a spindle on the x-y axis.


I have not settled on a PC platform yet. But concidering making it simple and cheap as a start. And I can upgrade if i find myself in a situation where it is meaningful to throw more money at the platform.

Questions:
1. What are my simple I/O options, a RPI5 and a basic I/O card, a BeagleBone with a cape ? Any recommendations ?
2. If i later on do the A axis i plan on installing the Sevo drive with a gearing that will get up to max 600rpm. . It is a small hobby lathe and  currently is has a 700W motor. Anyone with experience running servo on an A-axis - the servo motor i have is a 1200W type?

Hope this is meaningful, if not let me know and i will try to elaborate.

Thanks in advance!










 

there is added filtering and influence of real active ferror from feedback but I am not using it because it creating frequencies ... 

rodw wrote:

Very good stuff! I made a start today forum.linuxcnc.org/ethercat/58260-lichua...r-need-help?start=20
is dt = Delta time = servo period?
I'll have to follow you guys a bit more closely!
 

yes interpolation servo period.. Rod please check this repo  codeberg.org/endian/Beckhoff_AX5ZYZZ_lcec_driver here it is solved at all ... there is a .hal file example and everything needed 

Very good stuff! I made a start today forum.linuxcnc.org/ethercat/58260-lichua...r-need-help?start=20
is dt = Delta time = servo period?
I'll have to follow you guys a bit more closely!