Advanced Search

Fair point, and I get the frustration — I've felt it too. But I'd push back a little on lumping this in with the VC-backed subscription-everything crowd.
I've been at this for twelve years, solo. No investors, no team, no safety net. I own a small machine shop, I run Mac & Linux (hate Windows), and I built JetCad3 because the tools I needed either didn't exist, didn't run on Linux, or cost more than the machines I was running them on. FreeCAD is genuinely good software and I have nothing but respect for it — but my scope is much larger, and I don't think the end goal is achievable as an open source project.
The subscription isn't the goal — developers are. I've written my own geometry libraries, my own NFP and Skyline nesting algorithms — I own all of it specifically so I'm not paying upstream licensing fees that force my prices up. I want this as cheap as humanly possible. My end goal is to put the middle finger up at Autodesk, not replicate their business model.
I'm blue collar as it gets. I already run two businesses and support my family — JetCad3 isn't a get-rich scheme, it's a twelve-year bet that someone should build serious CAD/CAM that runs on Linux and doesn't cost small shops an arm and a leg. If revenue ever lets me hire developers who've actually run a machine, I think I can genuinely compete with Autodesk, MasterCAM, Dassault, and ProNest — and that's a win for this industry, not a betrayal of it. I want to be so much cheaper than them with more/high quality features that it forces them to lower prices when they start hemorrhaging subscribers...
One thing I maybe didn't make clear enough in my original post: there will always be a generous free/hobby tier for every workspace in JetCad3. Guys building their automotive project in the garage will always be a priority for me — because that IS ME.

Thank you all

Currently I am only running first tests on my machine. But mostly in sim. Still a lot of bug fixing to do. I was recently working on some performance improvements on the development branch which I will merge to main soon.

To give you more technical background: this controller is the result of over 10 years of experience in industrial machine retrofits.

The project was accelerated by the massive Mesa card shortage in 2022. Instead of waiting for hardware that wasn't available, we decided to develop our own high-performance industrial-grade solution tailored specifically for complex retrofits. This isn't just a prototype—these units are already integrated and running 24/7 in machining centers, proving their reliability in real production environments.

Key Hardware Specifications:

Core: Custom FPGA running HostMot2 firmware for rock-solid real-time performance.

Fully Isolated Analog Servo Control: Precise ±10V analog outputs with a 7V/μs slew rate. The analog outputs are fully galvanically isolated from the logic to prevent noise and ground loops.

High-Speed Feedback: Isolated encoder inputs supporting up to 10MHz—ready for any high-resolution industrial encoder.

Industrial Isolation & Protection: Complete galvanic isolation between logic and I/O. The 24V field power is protected, and the logic power is fully isolated.

Onboard Processing: Integrated Raspberry Pi Compute Module 5 (CM5), providing a massive performance boost for LinuxCNC and NativeCAM.

Thermal Design: Fanless industrial enclosure. No moving parts, no dust intake—designed to last in woodworking shops.

The goal was to create a controller that is fast, precise, and, most importantly, available for professional retrofits when the industry needs it most.

Hello. Could you tell me how to connect the old model centerfinder vers.ge/ru/blog/archive/vers-wl-v31 to the MESA 7I76 in the 5i25 and 7i 85s configurations? I supply the 7I76 with 5 and 24 volts, but the Vers WL v.3.1 requires 5-12 volts. Can I reduce the voltage from 24 to 10 volts using the XL4015E unit?
The Vers WL v.3.1 worked well on the MACH3, but it only had a 5 volt supply.
Thank you.

I was able to run 1.5.2 ribalda built on debian 13 on Papagno's hardware, OP are quick, and no perceptible grinding on a simple bench setup, with 0.9.5 linuxcnc-ehtercat sittner, This is not even the good PC hardware, it had 400ns of jitter, and the motors seemed fine, we will need to test better on the machine.
This was mostly to rule out kernel jitter, biggest culprit for slow OP is definitely ethercat 1.6.9
We (me and YangYang) have a refreshed 1.6.9 in the works, but not sure when that will land, squashing some bugs still, the code bases 1.5.2 riblada and 1.6.9 igh went in two separate directions, and they are hard to merge, will put best efforts in, then we will have a version for ourselves we can maintain ourselves, since igh is kinda of unresponsive and or uninterested, no clue...

jjdege wrote:

Hi,
What you're doing is fantastic. The installation is going well.
The simulations show great potential. Congratulations!
If I may make a point, looking at the video, I think it's better if the horizontal drilling unit
descends when it reaches its position. In the video, you can see it descending and then reaching its position.

That's just my opinion.

Thanks and congratulations again! 

Hi,

Thank you for the kind words and for the sharp observation!

Regarding the horizontal drilling unit: you are absolutely right. In the video, the Z and X/Y axes are moving simultaneously for speed. However, NativeCAM allows full control over the sequence.

In the final setup, we can easily configure the G-code to ensure the unit reaches its target position before starting the descent. Safety and precision are the priorities, and this is exactly why LinuxCNC + NativeCAM is so powerful—we can tweak these movements to fit the specific mechanics of the machine perfectly.

Thanks again for the feedback, it’s very much appreciated!"
 

Hello,

I'm trying to connect a Novusun NVEM with an STM32F207 CPU to LinuxCNC on a small PC (not a Raspberry Pi) for a mini lathe – just the X and Z axes, no limit switches, etc. Very simple.
I'm using just two TB6600 stepper drivers with 1.8 degree and 1.8A stepper motors, connected according to the many wiring diagrams I've found online.
So, Pulse+ and Dir+ are connected together to COM+ on the NVEM, and then Pulse- and Dir- are connected to the corresponding NVEM pins.
Everything is powered with 24V, including the stepper drivers and the NVEM.
All the LEDs are lit, so everything is receiving power.
What I've done so far:

1. Installed LinuxCNC on the PC
2. Flashed the Remora firmware for the NVEM
>> Source DOC: remora-docs.readthedocs.io/en/latest/har...ml#controller-boards
>> Firmware for STM32F207:
Either this one: github.com/scottalford75/Remora-NVEM/tree/main/Firmware
or this one: github.com/scottalford75/Remora-NVEM/tre...-NVEM-STM32/Firmware
I flashed the latter onto the NVEM.

3. LinuxCNC Ethernet configured on 10.10.10.100/24
>> Ping to the NVEM is running

4. As per the source DOC, the following steps were performed to get Remora from GitHub onto the PC:>> I noticed that some things were missing, including the basic configuration file "remora-nvem-basic" and the component file "Remora-nv".
I then copied these files to the appropriate folders on my PC.

5. Installing the components with halcompile:>> It was noticed that "Remora-nv" was missing, so I installed it again:6. I copied the example configurations into the LinuxCNC folder, along with "remora-nvem-basic".
>> After starting LinuxCNC, I can select various Remora configurations to start.

7. Uploading the configuration to the NVEM - for this, I had to install TFTP:- via a workaround involving the installation of pip3, etc., since some things are not installed by default in LinuxCNC.
>> When I run the upload script, I'm in the same folder as the simple config.txt file from "remora-nvem-basic":>> This seems to be working so far.

8. Starting LinuxCNC and selecting NVEM Basic, which should match the Config.txt file. Everything starts without an error message.
>> From here, I can cancel the eStop in LinuxCNC and also turn on the machine. But as soon as I try to move an X axis, I get "Joint 0 following Error."


I think it's not sending any pulses; at least nothing is reaching the drivers.
Although I assume it's only controlling stepper motors in the config without encoders—how does it know the following errors? It must be a configuration problem, right?

Does anyone know where I can check this?

Regards,

Hi everyone,Just wanted to share a quick look at a Mobirelli Partner 430 machine running with the CNC PROTON controller (the orange unit).The system is fully integrated with LinuxCNC, focusing on high-speed real-time performance and stable motion control tailored for this hardware. Our controllers are production-ready, featuring custom front I/O board configurations for Mobirelli, Biesse Rover, Masterwood, and others. It’s designed as a "connect, configure, and go" solution for industrial retrofits.Video of the setup:

• Horizontal drilling: youtube.com/shorts/eXKx0J_iorY
• Machine in motion: youtube.com/shorts/yrjaTZI8PxE

I would love to hear your thoughts or answer any technical questions regarding the integration!

Hi,
What you're doing is fantastic. The installation is going well.
The simulations show great potential. Congratulations!
If I may make a point, looking at the video, I think it's better if the horizontal drilling unit
descends when it reaches its position. In the video, you can see it descending and then reaching its position.

That's just my opinion.

Thanks and congratulations again! 

Hi everyone,

I’ve just uploaded a short demonstration of the NativeCAM (Python 3 / GTK3) port running on an industrial CNC machine. In this video, you can see it handling horizontal drilling operations smoothly.

The port is stable and performing well in a real workshop environment on LinuxCNC 2.9 / Debian 13.

Live demonstration: www.youtube.com/shorts/RL7QDfig8vQ

You can find the updated code here:
github.com/cnc-proton/nativecam-py3-gtk3

Feedback and testing are welcome!

Sir, I think that site has not been working for more than a year. If you need old pages or PDFs, those URLs may help you.

web.archive.org/web/20260000000000*/https://sa-cnc.com
github.com/Tpj-root/Linuxcnc_Kinematics/tree/main/Back_Up

Hello, i try the source you describe sa-cnc.com but none of these are available
Maybe something is changed.

Is new information available

Thank you in advance

Difference between 00 and 01 is bug at master/slave layer... If you will map everything at 00 it will works just few IDNs and main number of then will produce mess .. i do not know why it is that but its the fact ..

Second bug what many year ago was reported by @bkt or other skillfull user was that not supporting stuff mapping pdo is present and and mailbox is not changing... But it was fake data because main volume this ax5 have mapping pdos ON if an EEPROM does not tell you else....its just forgoten message from early days

I think its correct you can add other slave to your TC3 project.. give them naming with string length less than 16 chars and it will works well

Post command + output of you situation then here I can check it..

Thankfully we have that translator and its save anybody non nerdy in lcec one half life.. because I lost it too
 

At least there is something free with the added explanation on how to maximize the use of that free version, so thank you for that.

First post by new members must be approved by an Admin or Moderator, to prevent spam.