[time-nuts] Near-perfect chip for Loran-C frequency receiver

[Bruce Griffiths]

Carl Walker wrote:
> On Fri, 2008-07-04 at 01:02 +0200, Magnus Danielson wrote:
>   
>>> The analog side would need to allow for those signals also then.
>>>       
>> Naturally. The antenna-amplifier design will need to be more wideband
>> oriented. Should not be too hard thought.
>>
>> Cheers,
>> Magnus
>>     
>
> I respectfully disagree; while making the wide-band receiver is an easy
> task, you now have a family of unrelated signals - often of widely
> varying signal strength. Once any of those signals becomes large enough
> to drive the receiver non-linear, you rapidly run into issues.
>
> AGC to maintain linearity isn't practical in this case - since reducing
> overall receiver gain to compensate for one large signal - like WWVB if
> you're close by that one transmitter - will potentially drive down the
> gain for desired LORAN and other signals to the point where you can't
> acquire and track many of the weaker but never the less desired signals.
>
> There's more than meets the eye initially when you attempt a receiver
> design of this type - at least as far as the analog section goes. Once
> it becomes 1's and 0's it's all straight forward - at least as far as
> this old ex-analog guy is concerned ;-)
>
> -Carl
>   
This is only true if one is attempting to simultaneously track all the 
signals with a single receiver.
This isnt practical unless one uses separate analog receivers (each 
using AGC) dedicated to each signal to boost all signals to a similar 
level before sampling them with the ADC.

In practice, it may be simpler to dedicate one digital receiver to each 
signal  with the analog input filtering selected to suit a particular 
signal and location.
There's little point in dedicating resources to receive signals that in 
particular locations are too weak to be useful.

The 12 bit ADC's dynamic range will limit the disparity in signal levels 
that can be handled simultaneously long before the analog circuitry 
causes problems (at least in the case where the analog circuitry isnt 
driven into compression).


Bruce