LinuxCNC-RIO - RealtimeIO for LinuxCNC based on FPGA (ICE40 / ECP5)

Sometimes it can be an advantage to be a bit lazy
Otherwise, I wouldn’t have been able to reproduce it.
At first, I had the driver connected directly to the TangNano9k (3.3V),
then I had a quick Google to see if anyone else was having problems with the DM542. Yes, they were, because of the 3.3V. Now I’ve connected it to the Tangoboard with 5V signals and everything’s working perfectly, with the default settings

www.youtube.com/shorts/3IeqAu58WdI

Is your DM542 input pin connected like this?
 

My DM542 input pins are connected like this. I've used this DM542 driver on approximately 1000 units. I've used it on Linuxcnc LPT, Mesa 7i92, and Grbl 1.1 controllers without encountering any positioning error issues.
 

From what I see in your video, the connection is like this: PUL+ and DIR+ are not connected.
 

If I replace the DIR PC817 optocoupler with a high-speed 6N137, it can correctly return to the zero position.   
However, using grbl and mesa 7i92 does not require modifying the DIR pc817.
I may need to use a logic analyzer to observe the timing diagrams of the Tang Nano 9k output pulses and DIR
I'm from Taipei, Taiwan, and my English isn't very good.

 

viewsat wrote:

If I replace the DIR PC817 optocoupler with a high-speed 6N137, it can correctly return to the zero position. 

This is wrong way.
By modifying the driver, you risk damaging it and reducing its reliability.
Although the PC817 has a very low bandwidth, the correct solution is to increase the dirhold.