understanding AGC

[Greg Troxel]

With rtl-sdr dongles, there is AGC, and manual gain control is so messy
that AGC is the one true way. Please tell me if I'm confused and how
so.

I am guessing that the AGC attack time is fast enough that when a
preamble starts gain is reduced to have the signal not be clipping
pretty fast, so it can be decoded. And I am guessing the AGC decay time
is slow compared to transmissions so it doesn't come back up much
between bursts.

Because of AGC, signal levels of substantially all transmissions are
close to 0, often -0.5 dB. A signal level lower than that is a clue
that receiver can't ramp gain up enough to put the signal at full scale.

The noise level is an estimate of channel energy during time periods
when the decoder (or really the pulse recognizer) thinks that there is
no signal. This is either broad fuzz from vast numbers of devices or
thermal noise


Does anybody have info on attack/decay speeds?

[Greg Troxel]

(First, I am thinking about the situation where one is listening on some
frequency for a number of devices and thus expecting a wide variety of
signal levels. I realize some people want to listen to one specific
emitter and can figure out the level and adapt to it. However, I have
been seeing significant varatiion in snr from an outside soil moisture
sensor; the snr plummets when it rains hard.)

I am using multiple rtl-sdr dongles, e.g.
NooElec Mini and Mini2
NooELEC SmarTEE v2
RTL-SDR Blog v3

Certainly, with a better SDR that has manual gain control in reasonable
steps (and maybe even calibrated gain), and much more than 8 bits, it
can become reasonable to operate rtl_433 in manual gain mode. But for
the hardware I have, so far it seems that using AGC is the only
plausible approach.

Thus I wonder for the class of good-quality rtl-sdr dongles (basically
nooelec models with tcxo and Blog models), if anyone knows what the
actual attack and decay time constants are (or if they aren't linear;
there's no need in the digital world, compared to the early days of SSB
when every tube was a big deal).


Are you using manual gain control with rtl-sdr dongles? If so, can you
describe how you decided what to do, and how it deals with both strong
and weak signals from varying emitters?