RSSI and Smooth Attenuation

[NZ0I]

The attenuator that has been added to the design can be stepped in 0.5 dB increments between -0.5 dB and -31.5 dB. I believe that should let us smoothly adjust the receiver gain (the gain between the antenna and the attenuator) over that range.

We also have confirmed some ways to attenuate the SA605 by adding some external bias to the clock inputs, or to the RF input. (That's one positive thing about having left out the DC-blocking caps initially.) We should be able to accomplish the attenuating bias using an I2C pot, giving us (128 or 256) small attenuation steps that we can apply at the IF level. Again, I think we will be able to make the attenuation steps small and smooth.

I imagine the algorithm for applying attenuation to be something like this, starting from a very weak signal where we apply minimum attenuation and maximum gain at all the stages:

As the signal level rises from the noise it eventually reaches a threshold where the received signal level is high enough that the SNR is well above the receiver's inherent noise level, and we achieve something like "full quieting" in FM terminology. From that point on, we begin increasing the attenuator gradually in small steps, to achieve a very slowly-decaying AGC function, but with a faster attack, so that a suddenly increasing signal strength causes a rapid ramp-up of attenuation (gain drop). We want the decay response to be so slow that it changes very little over the amount of time it would take to swing an antenna 360 degrees. So if you are taking bearings, the attenuation ramps up (gain ramps down) so you hear a strong signal in the peak direction, and an easily-discerned drop-off as the antenna turns away from the peak direction. After a fast atten/gain attack response (as the received signal increases) the atten/gain should remain constant as the antenna is turned beyond the peak direction.

As we approach a transmitter, the received strength should reach a point where the receiver has dialled in 20-30dB of attenuation. At that point the LNA can be turned off, dropping the gain by ~20dB. At the same time that the LNA is turned off, the attenuator can be adjusted to reduce attenuation by 20dB. So the user would discern very little change as a result of the LNA gain being lost.

As we continue to approach even closer to the transmitter, the LNA remains off, more attenuation is dialed in until 20-30dB of attenuation has been added. At that point the gain stage (or one of the gain stages) can be turned off, dropping the gain. At the same time that the gain stage is turned off, the attenuator can be adjusted to reduce attenuation by the same amount. So the user would discern very little change as a result of the gain having been lowered.

As we get even closer to the transmitter, eventually all of the gain stages have been shut down, and the attenuator has been dialed up to its full 30dB of attenuation. At that point, the gain of the SA605 is dialled back incrementally, until the receiver is almost totally deaf. Hopefully we can achieve the equivalent of >120dB of smooth attenuation in this manner.

As we move farther from a transmitter, the process is reversed, until full gain and minimum attenuation are applied with smooth transitions over the full range of settings. (We will also need to devise a way for the receiver to quickly increase its sensitivity in response to a weaker fox coming on the air. This might be a manual process, accomplished by pressing a "trigger finger" button, or it might be automated by synchronization with the fox cycle.)

If this all works, and we devise an intuitive way to convey the current attenuation level to the user, the user will always have a very clear idea of how far he/she is from the transmitter. The attenuation level could even be calibrated (during ARDF practice sessions) to provide estimated distance to transmitter (in meters/feet) that can be displayed on the LCD.

We have several options for conveying the attenuation setting to the user via the headphones. The processor-generated tone can change in pitch, and/or in volume. The Russian 80m receiver I've played with did that. At weak signal levels (zero attenuation, max gain) the receiver would generate no tone at all. Somewhere around "full quieting" (for lack of a better term) a high pitch signal would become discernable. As the attenuation knob gets turned to lower attenuation settings (the Russian receiver doesn't apply attenuation automatically) the pitch would decrease. The pitch would continue to drop until, only a few meters from the fox, the tone is about 50 Hz and booming into the headphones. The tone actually got quite annoying, but since one is spending little time right next to a fox, it was bearable.

The tone would turn off/on in sync with the received CW signal: I think that was probably done by detecting PLL lock for a PLL tied to the audio of the received signal. I suspect a PLL was used because the tone would disappear if you tuned too far away from the frequency of the signal (beyond lock), even if you could still hear the received signal in the headphones.

We might be able to accomplish the same thing, or perhaps do even better. The current receiver design includes a sensitive RF detector with its RF input attached just after the BPF in the front end. Although it is a sensitive RF detector, it is still relatively deaf in comparison to a sensitive receiver. So when the receiver attenuation level is high, and a signal is clearly discernable to the RF detector (and it rises and falls in sync with the RSSI?) it will be a clear indication of close proximity to a fox. The receiver might utilize that information to improve the confidence of close proximity, and it might synchronize the tone output (on/off) to coincide with the detector's signal indication.

All this reminds me that I should include analog switches in the design to allow the 2m and 80m receiver sections to share the RF detector and Attenuator devices.