[RTTY] K3 reduced-bandwidth RTTY analysis

[aflowers at frontiernet.net]

> The one fault I found with Andy's AFSK work is that he used the K3 with AFSK filter 

Just to clear up any confusion, the *original* article that looks at the effect of waveshaping shows the K3 AFSK TX filter in only one of the figures (Figure 15 to be exact).  All the others do not have this filter engaged.  Please keep in mind that until very recently, the only examples of bandwidth-reduced RTTY were from AFSK transmitters, and that allowed me to use real evidence both on the bench and off the air to illustrate the possibilities.  Without that I only could have spewed forth some hypothetical abstractions about how to improve the QRM situation...I'm sure that would have made for a thrilling discussion on this list....One of the main points was that there is no reason bandwidth reduction couldn't be done using logical keying internal to the radio so that we can forego a mess of cables, boxes, and gain knobs that are sometimes employed for AFSK.  Maybe Elecraft will be the first to make that available.
For any latecomers to the party, here is the original article which has been updated to include the fact that fldigi now has a shaped transmit spectrum by default.  I would also suggest that you read the email threads in January 2013 on this reflector's archive, as Joe and others bring up some points of agreement, dissagreement, and other things that I did not address.  If you take the time to read through my stuff, please read or at least skim through the reflector archives as well.
http://www.frontiernet.net/~aflowers/k3rtty/k3rtty.html
 
Some some of the more active email threads to get you started:
 
http://lists.contesting.com/archives//html/RTTY/2013-01/msg00134.html
http://lists.contesting.com/archives//html/RTTY/2013-01/msg00170.html
http://lists.contesting.com/archives//html/RTTY/2013-01/msg00224.html
Here are some real live spectra from the 2013 RTTY roundup so that one can get and idea of the variance that is out there in the wild (warning, signal #3 is PG-13):
http://www.frontiernet.net/~aflowers/rtty_examples/

Joe, I think you are referring to the follow-up article that looks specifically at the effect of the DSP pre-processing ahead of the RF exciter, i.e., the AFSK TX filter in the K3.   The idea there is just to demonstrate how effective (or not) that strategy really is even with imperfect, or even truely broken transmit chains.  Now, this happens to be a K3 with features that Elecraft implemented specifically in that radio, but there is no reason that DSP pre-processing couldn't be employed in other radios, should the manufacturer choose to.  An FT-1000D will not behave like this under the abuse I subjected this radio to, and there is a bold disclaimer that says as much before the implements of torture are brought out.  That's important if you are thinking about changing the way you do things.
Here is that article:
http://www.frontiernet.net/~aflowers/k3_afsk/
Of course, if Elecraft goes through with the wavshaping in their FSK generator one advantage to using AFSK with this particular radio will go away.
 
Andy K0SM/2