I've received the DA-3605 module (4 channels, +/-10V). Couldn't pass it up — it was brand new, sealed, and cost me only $23. Now, to complete the setup, I just need to find the encoder modules. While I'm looking for them, I want to try running a combined Mesa + EtherCAT setup. 

I tested an Intel i226-based network card from AliExpress on my Lenovo M710q, installed in place of the Wi-Fi module. Unfortunately, the connection was unstable, so I think I need to look for a different solution. 

Does anyone know of a proven, ready-to-go solution for adding a second network port to a Lenovo M710q?

Unfortunately, I only have one. The supplier has sent me another EtherCAT board for the MH300 to see if that is the source of the problem.
Delta has indicated that this is a bug; on other CNC systems, they use a flag to disable the Product ID check, which avoids this issue. Is there a way to do this in our XML file?

I have added some more functionality to the grid script it can now cut gridlines drill holes and make counterbores.
 

Script attached

Hi! yes, i have seen your work, nice, really nice but i tested it online and i still find the issues linuxcnc still has (the most annoying one is the single block; when moving press single block. when it halts, press 5 times cycle start, it will not step one block only but it will step 5 block!).
another thing: blending is based on geometry not on feed! so when you single block it will not stop exactly at programmed position but it will stop just before blending occurs (because blend is geometrical, it actually appends a circle movement to it!, just like siemens g642, no industrial cnc does that unless wanted to with specific code)
another thing: linuxcnc new scurve is based on ruckig (cruckig actually), great but not great at same time.. i have used its predecessor (reflexxes libraries) for a looooong time and they are meant to be used for path planning axis by axis, movement normally does not give it a straight line unless phase synch. in this case everything is working because path is considdered as one degree and then mapped to each axis, good idea but not best approach
classicladder: no comments ()
thats basically what i started to do, complete re-write of cnc from scratch to overcome all this issues i faced while trying to modify lcnc.

it would be nice to share more deep thinking about how lcnc structures and how it can be done better

regards!

Now the screenshots and the GitHub link is there.

OK thank you!

Problem solved!
I bought the board as a 7i85 and never double-checked the exact model. After investigating further, I discovered that it is actually a 7i85S, not a 7i85.
This explains why only one Smart Serial channel was working and why the RX measurements were different on the other channels.So there is no hardware fault, just the wrong board.

Thanks for your help!

Is this open source? Where to find on GitHub?

Yes, clearly: if I turn the machine off for 15 minutes and then turn it back on, the MH300’s product ID changes.
I have tested this—even without launching LCNC or running the EtherCAT XML command—and found that after the machine has been powered down for 10–15 minutes, it returns a different product ID. Although this ID is consistent each time it occurs, it differs from the one specified in the manufacturer’s XML file (which I used in our own XML file).
In practice, if the customer turns the machine on first thing in the morning—following an overnight shutdown—the machine fails to start; the MH300 remains in the "pre-op" state with a product ID of 10400000. However, if the customer then turns it off and back on again after 2 minutes—allowing enough time for the inverter’s capacitors to discharge—the machine starts up correctly, with the MH300 showing the product ID 104001F0 (the one found in the manufacturer’s [Delta] XML file).

Why can't I post screenshots here? 
So annoying.

This GUI is inspired by industry standard controllers and also probe_basic. Its goal is to provide a modern, graphically and feature rich GUI which is familiar to 
industry users and can be used in industry education, optimized for physical control with a CNC control panel and industry like custom CNC keyboard.

You can find it on GitHub:  Demo AppImage / Source code on GitHub

It is written from the ground up in Rust/egui and therefore starts up very quickly and has good performance even with syntax highlighting,
and context aware 3D display (highlight current line of code in the toolpath, loop iteration counter display, ...).

We just had our first release. It is now running on my machine and I start to find it useable now. Will be daily driving it while fixing issues.

Some features:

• G-Code syntax highlighting (works also with larger files)
• current G-Code line highlighted in the 3D view
• tooltable optionally integrates with loobric server ( github.com/loobric/ )
• automatically shuts down Internet / WIFI when a program is loaded or MDI to safeguard latency, brings it up again on program unload / MDI done
• Integrated G-Code help
• probe basic inspired probing
• Loop variable / Repeat identification and live monitoring (so you know where in the loop the program currently is)
• Support for lathe / canned cycles (currently only tested in the simulator)
• Integrated LinuxCNC log viewer and HAL browser
• mouse and touchscreen navigation also works where it makes sense, if available (3D view can do touch pan and pinch zoom), although not everything, lists can be touch scrolled. 
  

Main overview screen:


Program screen (note that the current selected G-Code line is highlighted):
 

Control panel simulation (this could be used on a touchscreen, although a physical panel is available and recommended):
 

MDI:
 

Probing:
 

Integrated interactive searchable offline G-Code help:
 

Integrated HAL browser:
 

Running on my machine: