Hello Johan and group,
About a PIC-based digitizer, clocked by a 10 MHz reference.. I was also
thinking along these lines, but I'd use a different microcontroller - at
least one with a 'real' serial port which would greatly simplify the
firmware. Possibly something with USB, or (even better) an Ethernet
connector, to stream out the data
via UDP/IP or TCP/IP.
you wrote:
> Do you think that a 8, 10 or 12-bit ADC followed by decimation in SL
would be good enough for mHz/uHz FFT resolution? Will strong sferics be
a problem or will some hard clipping be helpful as a "Jimi Hendrx mode"
impulse blanker if the input BW is selected with care? Maybe
quantization noise would become a problem with such low resolution?
<
Difficult to say. The impulsive noise from sferics is not such a good
'ditherer' as white gaussian noise.
The best way to find this out would be making a wave file recording with
8 bits only, from a synchronized soundcard or whatever, using 'real'
radio signals.
But I think without analog bandpass filtering a (standard) PIC isn't up
to the job. For example, I am only about 100 km away from the 23.4 kHz
transmitter, and that signal is incredibly strong here (I can see it on
an oscilloscope with a piece of wire, with only a simple RC highpass
filter at the scope's input).
The difference between DHO38 and the strongest 'dreamer's' signals seen
here (from Uwe) was in the range of 80 .. 90 dB, iirc.
Maybe time to get back into the ("hardware"-) DSP business and simply
try ;-)
73,
Wolf .
|
