Another 8x8 matrix based display...
This is not a real project, just a try to hide the LEDs for better WAF.
Print still needs some impovements, but from a little distance you can't see the details.

With radios from ITEAD this issue is obsolete. They seem to populate original nordic chips, or even good fakes.

[Solved]
Initial Problem were incoming messages at sensor node during reading an image from SD card (170ms). Increase of serial RX buffer size solved the problem.

Good to know:
The serial buffer size is not defined in:
c:\Program Files (x86)\Arduino\hardware\arduino\avr\cores\arduino\HardwareSerial.h
You can change the numbers with no effect to your sketch.

You have to change:
c:\Users...\AppData\Local\Arduino15\packages\arduino\hardware\avr\1.6.14\cores\arduino\HardwareSerial.h

Then it will work :smiley:

It's not working. FHEM sends messages instantly and does not wait until a node wakes up.
New features unfortunately not supported by FHEM.

The splash screen at startup is not redirected to a new serial port, if MY_DEBUGDEVICE is used. It still comes out of hardware serial(1). Tested on NANO/ProMini with latest mysensor lib. All other debug information comes out of the new defined MY_DEBUGDEVICE.

I count in every node, if send() returns false and send the number to controller to get an idea about rf quality.
I do not retry, because Mysensors already retries, right?
I use NFR24 and RFM69. Behavior is sometimes strange. No NACKs for weeks and then a really high number of NACKs for a few days. Setup not changed. I have no idea why... Same for indoor and outdoor sensors.

I ordered modules from ITEAD. They seem to have real nordic chips populated. Each module comes in paper box and ESD bag. The modules look like the cheap ebay stuff, but with 0402 components.
With these modules I got rid of my issue with repeated messages.
(https://forum.mysensors.org/topic/5588/multiple-messages-with-same-content-received)

Most of these cheap boards do not have any hysteresis and tend to oscillate during output switching.
Populate a resistor (100k...330k) for hysteresis and remove a capacitor. I my case this did help.
Unfortunately in German, see the picture:
https://forum.fhem.de/index.php/topic,54408.msg459962.html#msg459962

I've opened an issue for that.
https://github.com/mysensors/MySensors/issues/1452

According to API payload is same on ECHO message.
I feel this should be handled/fixed in lib, but I have no idea how to do...
If it is not possible to fix in lib, it shall not be allowed to send message with no payload anymore, which controller has to make sure then.
@rejoe2
Regarding FHEM, at the moment we should keep it as it is. No payload is not the normal use case.
I flashed gateway with 2.2.0 and this fixed the problem for me.

DualOptiboot: if you do your own board layout, always keep the CS signal of the external flash at pin D8. You can change the pin number in your sketch, but not for the bootloader It is hard coded in the bootloader!
In this case the uploaded firmware is written to external flash by the sketch, but the bootloader can not read it and not write it to internal flash. Firmware will never get updated. Took me some time to find out...

Yes, no real EERPOM, but emulation in flash. EEPROM failure was the first finding when checking for this error message. Maybe somthing else...
Now complete sketch is ported from AVR ProMini to STM32 and is runnig fine :+1:

@virtualmkr thanks for pointing to _serialProcess(). This help to understand a little more how it works.
I finally found the issue: bias resistors (10k) at A and B lost connection, so during bus idle RX was low and not high. For AVR this was not a problem, but STM32 did not synchronize right to the frame.

No. For RFM69 you need Sensbender bootloader and external flash. Then it works.