Thank you for the information, I sent an email. When I built my DIY board, I came across the same problem. It was caused by one component, think the MPU, which was not connected to 5V voltage. So I think if one of the i2c components has a problem, the whole bus is not working. This will, of course, not help if the components are surface mounted and can't be checked one after another.

To be honest, there is only question here: is there a chance to get the NyBoard replaced or not? It seems to be broken, at least the i2c part. What caused it? I don't know, maybe the switch, maybe my soldering, maybe the PCB is broken. If there is no chance of replacement, I will stick with my DIY board (which works) and try to improve the installation to be mechanically more stable.  I invested so much time in trying, checking and soldering/de-soldering that I suspect this to be hardware problem. I learned a lot from this kind of debugging, but I would refrain from investing more time on this topic if there is no chance of getting a replacement.

Will do when I'm back home. Tried some things yesterday: replaced the switch with a spare part, it does not work. Checked SCL/SDA connection to ground/5V. Not shorted, Pullups between SCL/SDA and ground are ok (680 Ohm). Soldered the Raspberry Pi connector and connected it to an M5 stack ESP32 with i2c scanner running and switch set to RPi, but it shows the same effect.

It does not find anything and never finishes. In fact I modified the script to show the iterations and it does not even start, scanning hangs at iteration 1.