Tested different Arduin's

After days of struggling to get a simple IoT sensor (NRF24L01+ with
the door/window sketch) working on the IoT platform I finally have a
stable environment. Maybe the next could be helpful for others:

In the webclient under the settings for MySensor driver the default
channel is 90. In the file MyConfig.h (in the MySensors library)
however the default channel is 95. I both changed them to 76 (both 95
will work too).

I tried a brand new 3.3v Arduino Pro Mini with the door/window sketch.
The Mini was powered with 4 AA batteries (total 6 volt) and a LDO
3.3v regulator. The NRF24L01+ has a 10uF cap across VCC and GND.
Communication between the Mini and the IoT server was very buggy. A
lot of errors.

I than tried a Nano powered by a PC USB port. Perfect. No glitches.
Next was a 5V Pro Mini without regulator and power led (powered by the
same 3.3v LDO regulator). Once again perfect and no glitches. So in
compared to the 3.3V Mini the 5V version had no power led and
regulator so I removed these also from the 3.3v Pro Mini. Tried the
same sketch again and voila, perfect and no glitches.
I think the NRF24L01 needs as much juice as possible and removing the
power led and regulator is needed to get a stable situation.

I’ve tested this for a couple of hours and there are no more error
messages in the server log file. Perfect!

wait so are you powering a 3.3v MCU with 6v...?

Because I don't have an enclosure for 3 AA's

I think you may damage it in the long term like this, do you have one for two??

You could easily DIY something for it...
www.instructables.com/id/DIY-AA-Battery-Holder/

No. Like I said above I use a 3.3 voltage regulator. The voltage regulator is connected to 6v. So not 6v directly to the Mini. In that case there would be smoke I think.

Just to clarify: he is using an external voltage regulator to connect the four batteries to the VCC (not RAW) of his arduino, and in turn the onboard voltage regulator of the arduino was de-soldered (since it is not in use anyway).

Another idea: what about powering the NRF24 module directly from your external regulator at 3,3V, and power the arduino with the un-regulated 6V, using its onboard regulator? This way, no de-soldering would be needed. Since the NRF24 module doesn't draw power from the arduino, it cannot affect the stability of it, and the module can draw as much current as the external regulator can deliver. The arduino will most definitely work pretty stable when running on its own regulator, provided it does not have to feed the NRF24 module.


regards