Problem with battery supply between two Arduinos that uses serial communication

I have a project that uses two Arduinos, the first is in charge of communication between Arduino and LoRa module (sx1276), the second is in charge of getting some values of GPS, and temperature sensor and these Arduinos are communicated with serial communication. All code and hardware works perfectly until I remove wire supply from my PC to Nanos and add a battery; the battery characteristics are 18650 ; Li-Ion; 9900mAh; 3.7V I use different combinations to supply all the circuit:

  • I uses two batteries (7.4V) to supply all circuit, as you can see on the image. it doesn’t work
  • I uses one battery (3.7V) to supply only the LoRa module sx1276 and connected USB cable to supply the Nanos. It works
  • I uses three batteries, one to supply LoRa module, and the others to supply one Nano, the last Nano is connected to USB cable. It doesn’t work

I use the command Serial.print() to assess the received data on the Arduino with LoRa module, and in the majority of cases the data is corrupted, i don’t understand why battery supply doesn’t work if the current consumption is okay, and all voltages that I use are okay … 3.3V - 5V on the Arduinos, sensors and LoRa module.

I think that is a noise problem associated with GND, but I can’t fix it.

Thanks ind advice

Here is the problem with Fritzing. Right away I look to see how the power is connected. But there are no labels on the Nano (or they're so tiny I can barely make them out). So instead of loading up my Nano pin map just to do that, I leave the problem for people with more time than me.

But... it looks okay. Did you triple check your wiring?

What do Yo mean by I remove wire supply from my PC . Personally I use USB cable and their connectors when powering from the Pc, not "wires".

You write "it doesn't work". What do You mean by that? There are 2 nanos, GPS, Lora.... Are they all completely dead?

As @aarg says, Fritzíng diagram looks nice but are more ore less useless.