Coordinator node locking up

Dear Digi

We appear to have reached the limits of your device - or are using it in a way not intended?

Here is the symptom. After some 2-3 hours of intensive, smoke testing operation ( coordinating repeat request/responses between 3 other radios with ~100byte packet sizes ), the radio goes into a state in which only a power cycle will recover it from.

In terms of hardware, we are running the Coordinator from a Raspberry Pi 3b over its high speed GPIO based UART using a nodejs library (no CTS/RTS).

The units are the surface mounted 900Mhz low powered SX modules certified for use in NZ, mounted on a breakout board to expose the standard XBee foot print, and then onto a WaveShare ARPI600 shield to access the GPIO.

They are configured as follows:

BD=7
AP=2
HP=0x01
D7=0x00
NH=0x01
CI=0x01
BH=1
NH=1
MR=0
MT=0
TO=0x40

My assertion is that we are somehow overloading the unit - but exactly how is unclear:

  • Serial Buffer over run
  • Excessive temperature
  • RF glitch or mesh networking glitch?

Tracing down exactly what is one thing, and rectifying it may be quite is another.

For field purposes our smoke tests are excessive, but naturally the failures cause some concern and demand of us that we attempt to resolve them - ideally by being able in advance to detect such failures and be able to avert them.

Welcome your thoughts

BTW. Great product!

You are most likely having a buffer over run since you are not using CTS flow control causing corrupt API frames.