search

ImperialSpaceSociety / picotracker-Lora

This is the code base and hardware design files for the Lora pico tracker to be flown around the world
13 stars 3 forks source link

The rx window timings may be off due to clock drift #28

Closed MedadRufus closed 3 years ago

MedadRufus commented 3 years ago

On ICSPACE19, for several days, the tracker failed to receive a packet from gateways on the ground to complete a join request. It made many attempts. I suspect its due to the timings of the RX window being off due to clock drift.

The clock source for this timing is a Low Speed External crystal which is not temperature compensating. The timing window needs to be accurate to within a few milliseconds.

One solution is to use the GPS 1 PPS clock input to caliberate the crystal.

MedadRufus commented 3 years ago

the clock needs to be at least stable to 1 parts per thousand. At most 1 millisecond off.

The RX window has to open within 1millisecond of the specified time.

The first rx window in happens 1 second after tx and the second window happens 2 seconds after tx

But looking at these numbers, it it simply has to have a frequency stablility of 1 parts per 2000. Even a normal crystal oscilator should be able to provide that. 500 parts per million.