[linrad] RE: Detecting phase modulation with undersampling

[Leif Asbrink <leif.asbrink@xxxxxxxxxxxxxxxx>]

Hi David,

> I has anyone any experience of demodualting phase modulated signals?
> 
> I am tryiing to measure the phase shift between two signals, when they 
> are derrived from the same source? I guess you could say this is a form 
> of phase modulation, but whereas for normal 'PM' receiver you don't have 
> access to crystal developing the signal in the first place, I do in this 
> case.
I would use two schottky mixers with 90 degree phase shift between them
and sample at DC and audio frequencies. Presumably you can amplify
the reference signal easily so it would drive a pair of SBL1 or similar.

To calibrate, send a frequency shifted signal into the signal input.
It is easy to do the algebra to find out how to combine the two
signals so you get a good sideband suppression. (The usual direct
conversion stuff.) Then you have an accurate phase meter.

This idea would be useful only if the ~200 MHz signal does not change
its center frequency too much because the calibration will have
a limited bandwidth. Depends on the accuracy you need of course,
but presumably you could cover 25% bandwidth if you calibrate at every 5%. 

> The problem is I need to undersample this. I have an A/D that can do 65 
> MS/s, but the input can be upt to 200 MHz - in pracitce I will use 70 
> MHz. However, since the bandwidth of the signal is narrow (< 1 kHz), I 
> am not breaking the Nyquist criteria, which states you need to sample at 
> no less than twice the *bandwidth* of the signal, and not twice the 
> frequency as most people seem to think.
But then you need two A/D converters. 50% of the time would be enough
so you can use only one and switch it between the reference and the 
modulated signal. 
 
> I then intend doing the processing on a dedicated DSP chip (TI 6701), 
> rather than a PC's CPU.
OK. I would use one of the dedicated chips to resample for a data speed
that is convenient to the PC. I think there are a few different types 
of boards becoming available that are intended for general purpose HF
SDR receivers. Just bypass the input filter, the A/Ds typically handle
200MHz without any problem.

73

Leif / SM5BSZ

 

LINRADDARNIL