A new RFM69 RSSI Mystery

[joelucid]

Today's insight: background noise RSSI measurements on RFM69 are path dependent! If you measure RSSI and during the previous RX period a packet was received, the RSSI measurements of background noise will be systematically about 3 dB lower than when no packet has been received.

For repeatable RSSI measurements switch the receiver to RX for 10 ms without receiving a packet, then to standby. RSSI measurements made during the next RX period will provide repeatable results.

Joe

[captcha]

Could that be the work of the AGC?

[joelucid]

Could that be the work of the AGC?

You'd think, wouldn't you. But I have AGC switched off and even tried with DAGC off. Still I can produce it reliably.

[WhiteHare]

Today's insight: background noise RSSI measurements on RFM69 are path dependent! If you measure RSSI and during the previous RX period a packet was received, the RSSI measurements of background noise will be systematically about 3 dB lower than when no packet has been received.

For repeatable RSSI measurements switch the receiver to RX for 10 ms without receiving a packet, then to standby. RSSI measurements made during the next RX period will provide repeatable results.

Joe

I wonder if it has anything to do with the optional improved sensitivity setting?  IIRC, it reduces apparent background noise by about 2 or 3 db.  Coincidence?

i.e. do you get the same lingering effect (see quotation) with it turned off as you do with it turned on?

It always seemed a bit strange to me that its use seemed to be reserved for just the one special case where the absolute greatest sensitivity was required.  In a different thread, I asked "Why not just leave it on all the time?" and the only answer offered was that it probably increased power consumption.  I don't think anyone ever quantified how much extra power it consumes, but it would be interesting to know.

[perky]

Leaving it on has a detrimental effect at the top end, you may gain 2-3dB at the bottom but it turns on a preamp I think and it saturates at high signal levels, I read somewhere you can lose 10dB.
Mark.

[joelucid]

Leaving it on has a detrimental effect at the top end, you may gain 2-3dB at the bottom but it turns on a preamp I think and it saturates at high signal levels, I read somewhere you can lose 10dB.
Mark.

I generally do have the sensitivity boost on. You can have two rfm69hw's with dipole antennas sending at maximum power right next to each other with lower baudrates without getting rx problems. And that's with AGC off and lna set to max gain.

It's only the higher baud rates (100kbit and up) which then get link errors.

[perky]

OK, thanks. I've just enabled it on my system (12k bps) with RFM69CWs and they seem fine right next to each other with dipoles.

I assume there's no great current hit with this option? Also, if it's an LNA on the front end presumably it reports a higher RSSI rather than dropping the noise floor, this appears to be the case from what I can see but it's difficult to tell.

Mark.

[WhiteHare]

IIRC, back when I did my long series of RSSI measurement experiments, turning on the sensitivity boost actually reduced  the noise floor as measured by RSSI.