[ntp:questions] Re: WWVB 60kHz Receiver

David L. Mills mills at udel.edu
Fri Apr 21 21:11:07 UTC 2006


I've been down the fancy decoding path myself, e.g., the WWV driver, 
which is a theoretically optimum linear receiver. However, much of the 
crud found at WWVB and DCF77 frequencies is bursty, which is what my 
LORAN-C receiver and program is good for. The Spectracom receiver is 
actually quite good: however, it was designed to cope with Gaussian 
noise, not suffer a 20-dB clobber by an interfering buzzsaw signal.

My precerred approach, should I accept the assignement, would be to tap 
onto the I and Q baseband signals in the radio, chop at something like 
10 kHz and feed to the L and R inputs on a sound card. I can take it 
from there. My problemis that the SNR has become so degraded that the 
very good PLL in the radio doesn't lock up.


Jonathan Buzzard wrote:

> On Tue, 18 Apr 2006 02:19:58 +0000, Danny Mayer wrote:
>>Jonathan Buzzard wrote:
>>>At the moment it exists as a Matlab program that runs against a previously
>>>recorded HBG signal, that I normally cannot decode reliably with the
>>>simplistic methods. However the Matlab program can decode the signal
>>>accurately better than 90% of the time. It will need converting into C and
>>>made to work in real time to be useful however.
>>Normally I'd suggesting running the MatLab compiler which can convert it
>>to a binary but the overhead can be horrendous. I've done it but I don't
>>think you'd be happy with the time it would need to run your algorithm.
> I would not be happy with the resultant code. Using Matlab is purely an
> algorithm development tool. Once I am happy with the algorithm I will
> rewrite it in C.
> JAB.

More information about the questions mailing list