On Aug 23, 2011, at 3:06 PM, Ed wrote:
> I may be missing something here, but how are the newer rigs generating
> FSK ? Are they shifting xtals or just sideband tones ?
Various methods.
Back in the FT-990 and FT-1000D days, Yaesu had used keyed-AFSK. The FSK
keying signal at the rear panel was used to key an internal AFSK generator,
which fed the SSB exciter. So, even though you "think" you are doing FSK with
those rigs, technically they are AFSK (which led to people calling them
"keyed-AFSK").
Funny that you didn't get any complaints about AFSK sidebands from the FT-1000D
and FT-990. Albeit, Yaesu forces the RF voice processor off when you select
"RTTY" and the AFSK generator has a fixed amplitude. The AFSK generator is
crystal controlled, so the shift at least stayed put like some of the FSK rigs
of that generation.
The Omni V and Omni VI were "direct FSK" rigs and the keying signal shifts an
RF oscillator with a trim cap. This frequency keyed RF is done at an
intermediate frequency and mixed to the operating frequency. Yes, there is
also an image, but presumably very low level and no inside a ham band. Because
of the fragile trim cap, you can often identify Omnis back then by how far off
their shift frequency is when the trim cap value changes over time, or over
temperature :-).
By the time of the FT-1000MP generation, Yaseu had changed to a more direct
version of FSK, by keying the direct digital synthesizer chip, which is how
many rigs from that generation do it. Many DDS chips come with a FSK and PSK
pin, and you can find them in RF synthesized generators such as the Novatech
409.
The rig designer cannot modify the FSK modulation from the DDS chip of course,
so the amount of keyclicks from an RTTY signal in the hands of the DDS
designer. The DDS typically has to allow at least 300 baud and perhaps even
higher baud rate -- thus, the keyclicks from most FSK rigs are wider than
needed for a 45.45 baud transmitter.
Modern rigs such as the K3 generates the FSK in its DSP stage. Mathematically,
it is not that different from "AFSK" but done at 15 kHz instead of at one or
two kHz with the more traditional AFSK. The image is thus further away than an
audio FSK signal. The DSP is probably (I don't know for sure, but there is
good technical reasons) also helping to suppress the image ("the other
sideband") further by the use of I-Q (in-phase and quadrature) techniques.
As such, the DSP rigs could waveshape the "afsk" signal, but I do not know if
any of them that does it. The only way to minimize RTTY keyclicks today is to
use AFSK with a software modem that implements RTTY wave shaping. The RTTY
wave shaping process is very similar to CW wave shaping that VE3NEA talks about
in his May/June 2006 article in QEX.
Software modems like cocoaModem actually does I-Q processing internally.
However, it has to reduce the I-Q signal into a real waveform to feed the audio
line input of the SSB transmitter. If the time ever comes when newer rigs will
accept I-Q inputs, you can generate really clean "AFSK" signals from software.
73
Chen, W7AY
_______________________________________________
RTTY mailing list
RTTY@contesting.com
http://lists.contesting.com/mailman/listinfo/rtty
|