Topband: BDR redux

[John Kaufmann]

on4kj at skynet.be wrote:

>   Two diodes back to back in front of a receiver is'nt that a bit of a
> simple?
>   My first thought in this case goes to a " Fourier series Delight ". Must
> be hard to correct by a audio filter.


There are a couple issues to consider with hard-limiters like back-to-back diodes.  The first is harmonic distortion.  For IF signals that are substantially sub-octave, i.e. occupy much less than one octave of bandwith, the distortion can be removed by bandpass filtering.  Decades ago, many engineering papers were written, all showing that hard-limiting of basically CW signals followed by bandpass filtering resulted in very little loss in SNR after the filter.  This was the basis for gain control systems in many digital data systems, using hard-limiting at IF.  

The second more serious issue is intermodulation distortion (IMD) between multiple signals in the passband.  The IMD products that fall within the passband of the filter cannot be filtered out.  The transmit speech processing systems used in many rigs today are based on IF limiting.  The SSB waveform is clipped and then passed through an IF filter.  As you crank up the processing gain, you can hear the distortion increase, due to rising IMD.  For moderate clipping, there is a net improvement in intelligibility over noisy channels because the clipping raises the average power in the signal.  For heavy clipping, however, the increase in distortion eventually begins to hurt intelligibility.  The same principles apply if clipping is employed in the receiver.  

The idea of back-to-back diodes at audio, followed by an audio bandpass filter is nonetheless a simple, but effective means of gain control, particularly for dealing with weak signals in the presence of large static crashes.  I've tried it and it does work pretty well.

73, John W1FV