Return to KLUBNL.PL main page

rsgb_lf_group
[Top] [All Lists]

Re: VLF: Clipping or blanking??

To: [email protected]
Subject: Re: VLF: Clipping or blanking??
From: Wolfgang Büscher <[email protected]>
Date: Fri, 21 May 2010 15:07:06 +0200
In-reply-to: <[email protected]>
References: <008f01caf784$779a11d0$0301a8c0@your91hoehfy9g> <[email protected]> <[email protected]> <[email protected]> <[email protected]> <[email protected]> <[email protected]> <[email protected]> <[email protected]>
Reply-to: [email protected]
Sender: [email protected]
User-agent: Thunderbird 2.0.0.24 (Windows/20100228)

Hi Peter, Stefan and the group,

Peter asked:
How to lock the sample rate of a sound card against a MSK modulated signal?
 The only idea I have is to demodulate it and locking a re-sampling process
against one of the two carriers - by software. May be there's the key.
Usually no resampling - that would be too much overhead. The soundcard's ever-changing sample rate is simply taken into account at various processing stages, most noteably the NCO (numerically controlled oscillator) which mixes the observed band down to baseband (before the decimator chain, and before the complex FFT). The MSK signal, when used as reference, is indeed demodulated (but not decoded of course) by squaring it after lowpass-filtering and decimation. The two peaks at f_center +/- 0.75 * bitrate are used to measure the instantaneous sampling rate (using a complex FFT, and comparing the phases in the peak FFTs beens between two calculations, i.e. once every few seconds). The program then tries to predict the sampling rate for the next chunks of digitized samples for the rest of the application, because the length of a processing chunk is usually much less than the measuring interval. Very fast sample rate drift cannot be compensated this way, but for FFT resolutions in the range of a mHz or less, it's sufficient.

Cheers,
 Wolf .


<Prev in Thread] Current Thread [Next in Thread>