For 170 Hz shift RTTY, a good AFSK signal (e.g., using fldigi and 2Tone in AFSK
today) is not much more than around 280 Hz wide all the way down to -80 dBc.
If RTTY signals are made much narrower than 280 Hz, reception will suffer. At
the same time, making a transmitted RTTY signal wider than 280 Hz does not
improve reception perceptively when received by a modern software modem that
uses a Raised Cosine Filter or variants of it (both fldigi and 2Tone use a
variant of the Raised Cosine filter).
You can see the analysis here:
http://www.w7ay.net/site/Technical/RTTY%20Transmit%20Filters/index.html
In practice, due to transmitter PA's 3rd order IMD, an RTTY signal will start
widening at around -40 dBc to -50 dBc level, instead of staying narrow all the
way down to -80 dBc.
On the other hand, even the best FSK signal will be much wider than what I
described above, as can be seen in the second figure here:
http://www.w7ay.net/site/Technical/RTTY%20Sidebands/sidebands.html
As can be seen, a perfect FSK signal is already about 800 Hz wide at -35 dB,
with keyclicks extending much wider.
The exception is the Elecraft K3 which applies filtering even when transmitting
FSK. The width of a K3's FSK signal is more like 350 Hz (so somewhere in
between a good AFSK signal and a good FSK signal).
(The K3 starts with a filtered analytic (I+Q) baseband signal, much like AFSK's
baseband signal. That is why it could be filtered. However, the shifting from
baseband to RF is done numerically in DSP instead of using analog product
modulators in more traditional superhet transceivers with AFSK.)
73
Chen, W7AY
_______________________________________________
RTTY mailing list
RTTY@contesting.com
http://lists.contesting.com/mailman/listinfo/rtty
|