Hi Chris --

It's really not that hard.  A long time ago I wrote a frequency measuring program for
DOS in C that measured the output of an anemometer.  It counted pulses
over a one-second period, then waited for the next full pulse, then divided the total number of pulses received
by the total time period over which they were received, to get
frequency.  This worked quite well and, using a microcontroller, was later made
into a product that would measure frequency up to 50kHz.

A microprocessor could pretty easily be made to do this same thing,
making tabulations 20 times per second of an input carrier signal at, say
1kHz.  The output could be via serial port or whatever.  To get 16-bit
resolution, 20Hz * 65536 = 1.3MHz clock rate on the microprocessor. You
would also need a real-time clock. Quite feasible, I believe.

One advantage of FM is that you can send the modulated carrier a long distance over
noisy lines and reject much of that noise at the other end.  The trick
here may be to design a modulator that has the noise level / stability
required.

Karl


Friday, May 04, 2001, 20:52:17, you wrote:

Cac>        Let me get this straight. You convert the signal to FM, you use the
Cac> A/D to produce a digital signal, you then analyse this for frequency?? You 
Cac> guys DO like hard work! I can't see this competing with direct A/D somehow.
 
Cac>        If you convert to FM, you can programme your UART to count pulses and 
Cac> measure the frequency directly, but this requires low level programming. 
Cac> Having got the frequency, you convert to amplitude... To get 16 bit accuracy, 
Cac> you need.... And the links to hamcom don't seems to work anyway.

Cac>        Regards,

Cac>        Chris Chapman

__________________________________________________________

Public Seismic Network Mailing List (PSN-L)