Hi Folks,
I have the spectrum of an RTTY signal posted at
http://homepage.mac.com/chen/.Public/RTTY/fskspec.pdf
This is an Acrobat (pdf) file.
A 45.45 baud, 170 Hz shift signal with 32 uniformly distributed
random bits is generated, and a Hamming windowed FFT is taken.
The FSK is generated in a phase continuous manner (the carrier
phase angle at the beginning of a bit is set to the phase angle
at the end of the previous bit).
[i.e., the FSK time waveform is continuous, although the derivative
of the waveform is not necessarily continuous]
There is no shaping of the data bits, data transitions are
instantaneous.
4,000 of these spectra are computed and averaged, equivalent
to something like 45 minutes of real time signals (but only took
a couple of seconds of computation on a current portable computer).
The vertical axis of the plot is on a square root amplitude scale.
This is so that the sidebands show up well. A linear power scale
won't show them at all, as you can imagine, and with a linear
scale, the details are not clear. A traditional log plot would
have overwhelming sidebands. So, I settled for square root
amplitude plot.
The tick marks on the horizontal scale are at 10 Hz intervals.
The vertical dotted lines indicate the 250 Hz bandwidth limits.
73
Chen, AA6TY
|