@cYnd Well I'm out of ideas. Could be a hardware issue, perhaps you shorted something when handling the module or you exposed it to 5V?
Is the VCC output from the Mini 5V or 3.3V? If you supply the mini board with 3.3V you should connect the module VCC to 'raw' if i remember correctly.
I use the RA-01 and RA-02 all the time so hopefully you will be very pleased with them. Good luck!
The 85 reading is actually an error message used by the ds18b20 to indicate it has not had time to complete the processing of the reading. Most commonly this is only seen when the node first boots up and can be ignored or fixed by adding a small delay. You will find a lot of information on this if you do a google search. I have never experienced it myself but from other posts on the net it seems power supply problems and long cable runs to the sensor can also cause the error. Some have also had success by using a different resistor value. Note you only need one resistor fitted near your nodemcu not a resistor at each ds18b20.
If the 85 reading is well outside your expected reading you could just filter out any readings over 84 as well.
@electrik It's highly unlikely. I have more than one type of module and more than one module of each type, none of them work. I also swapped the modules between working GW and Mega node and it still works, I swap the module between working GW and the node, it doesn't work.
yes this is possible with nrf52840, but unfortunately zigbee is not compatible with mysensors. Mysensors is a software stack like zigbee is.
I think you'll get more infos if you take a look on nordic semi forums. They have a zigbee stack for nrf52840.
@adfx Thanks for checking. I read once a blog/page about the use of multiple spi devices. It explained about the use of pull up resistors on the enable pin of every device you have on this bus for this specific problem. Just to be sure when I was testing it I did not connect the nrf24.
I added that resistor to this PCB. So far it's works with the MySbootloader and just burning the program with the USBASP. Even with the nrf connected.
I really believe that it was the chip itself. But I still have them waiting for a better test environment.