Change -12 dBFS to -6 dBFS AGC approach limit

[Abraxas3d]

Please set rx_hardwaregain via iio_attr on ARM to 12 or -12 (whichever convention is used). Continue to use slow_attack AGC.

Why Do This: MSK is constant-envelope. ADC can and arguably should run hotter than general purpose -12 dBFS, used for non-constant-envelope signals. -6 dBFS doubles ADC amplitude +/-594 to +/-1,026, improving all downstream SNR.

Verification: dbg_rx_samples_I_raw_out on ila_symbol_lock. We need to watch peak amplitude. Target +/-1,000+.

Known Repercussion: After changing AGC, all soft value thresholds will need review and recalibration. It's already in the HDL as a commented out block and documented in the header of frame_sync_detector_soft.vhd.

[MustBeArt]

I am pretty sure that changing the hardware gain won’t do anything to change the setpoint in slow_attack AGC mode. In that mode, the AGC computation sets the hardware gain and overrides anything we write into the hardwaregain attribute. This can be confirmed by writing to that register interactively and then reading it back.

So far I have not been able to find any simple procedure that just changes the AGC setpoint for slow_attack mode (or for any other mode besides manual). The AGC hardware in the AD936x is complex and powerful. The documentation seems to assume that you’re either going to use the default settings (and get the -12dBFS setpoint) or you’re going to take over the whole configuration of the AGC hardware and optimize it for your application. I frankly don’t know how to do the latter.

I guess it’s axiomatic that running the signal hotter is going to help with signal to noise ratio, but by how much does that help with performance?