Author Topic: Shortest preamble to generate RSSI for RFM69?  (Read 1110 times)

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Shortest preamble to generate RSSI for RFM69?
« on: October 27, 2016, 01:50:47 PM »
I've got a couple of questions for people:

1) What is the shortest preamble only that could be sent in order to guarantee a RSSI interrupt for a RFM69 receiver? Is it a matter of a few bit periods?
and
2) What is the power penalty in waking up a RFM69 from sleep, transmitting this short burst, and immediately going back to sleep again?

I'm wondering whether it's possible to lock on to a RSSI sequence even though there may be noise from RSSI events from other transmitters around. In analogue video for example you can lock on to a sync and extract it from noise even if it has been 'suppressed', i.e. DC level shifted into the video level during the horizontal blanking period. This is done by opening a gate to sample falling edges and locking a PLL to the edges it finds. I wonder whether a similar technique can be used for recovering a known RSSI sequence in hopping when locking a receiver to it.
« Last Edit: October 27, 2016, 07:44:09 PM by perky »

raggedyanne

  • Jr. Member
  • **
  • Posts: 80
  • Country: au
Re: Shortest preamble to generate RSSI for RFM69?
« Reply #1 on: October 28, 2016, 03:41:22 AM »
Poll the RSSI register periodically the valid RSSI value is from the received spike. Because of varying factors the shortest symbol length is two or total 6.25symbols but there is no guarantee of when the interrupt will fire after preamble detection.

joelucid

  • Hero Member
  • *****
  • Posts: 869
Re: Shortest preamble to generate RSSI for RFM69?
« Reply #2 on: October 28, 2016, 05:35:53 AM »
Mark, since rssi detection takes only 2 bits sending one byte would suffice for detection. You could go shorter by using continuous mode instead of packet mode.

There are a couple of startup costs from sleep that are outlined in the datasheet. The oscillator takes 250uS to wake up, ppl 80 uS to lock. Then there is paramp and a few other things but these are very short.

Wrt your idea to detect rssi within background noise: reminds me a bit of acquisition in direct sequence spread spectrum (LoRa etc). See http://www.wirelesscommunication.nl/reference/chaptr05/cdma/sync.htm for how it's done there.

What exactly are you trying to do?

Joe

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Shortest preamble to generate RSSI for RFM69?
« Reply #3 on: October 28, 2016, 06:51:22 AM »
Thanks Joe. I'm thinking about algorithms for achieving faster lock for a 50 channel frequency hopping system, and what the minimum information needed to achieve a lock might be. Specifically, whether superimposing a faster but far shorter sequence of carrier pulses in the pseudo-random sequence on top of the normal but slower packet sequence might work. This would meet the criteria of evenly spreading the channel usage and keep battery usage down. They're just ideas at the moment.
Mark.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1298
  • Country: us
Re: Shortest preamble to generate RSSI for RFM69?
« Reply #4 on: October 28, 2016, 02:07:48 PM »
From the datasheet, "The receiver stays in WAIT mode, until RssiValue exceeds RssiThreshold for two consecutive samples. "
« Last Edit: October 28, 2016, 02:15:19 PM by WhiteHare »