[RTTY] PSK31 is faster (Was FD RTTY Question)

[Kok Chen]

On Jun 27, 2012, at 12:47 PM, Alex Malyava wrote:

> Why don't we just invent/introduce some new RTTY standard -
> the one with 6 bits instead of 5 - covering whole alphabet and digits
> without any FIGS/LTRS and speed it up a little bit to compensate an extra
> bit?

There is no need to introduce another "mode du jour" even.

7-bit ASCII (CCITT ITA-5) RTTY has been FCC approved (see part 97.309(c)) for a long time now.  fldigi supports it, so does MultiPSK and cocoaModem, among others.

In a discussion (a year or even longer ago) on this reflector, I had shown that for most RTTY contest exchanges, ASCII RTTY beats out Baudot RTTY in speed, even when both are running 45.45 baud.  

You get rid of the FIGS/LTRS confusion (thus problem with USOS incompatibility either; USOS is a Baudot problem), allows lower case, and it still beats out Baudot in contesting speed.  It is when sending paragraphs of upper case text that Baudot wins over ASCII.

Because of the Teletype Models 33/35, the popular speeds for running ASCII RTTY was 110 baud.  At that speed, it will wipe the floor with Baudot RTTY.

> Or drop one stop bit to save the length? Or use 3 frequency FSK -
> shift left is "0", shift right is "1" and middle is sync/start/stop ?

3FSK may not be a good idea.  The reason is that the equalizer to compensate for selective fading will be at best very complex to build.  

2FSK has the very unique ability to fight selective fading with a very simple thresholding scheme.  Once you add more tones, you can no longer build simple ATC circuits.  

For that reason, you will find that there is nothing in MFSK16 (16 tones), DominoEX (18 tones) or Olivia that explicitly fixes the selective fade problem -- they all use long interleaved codes to fight QSB in general -- and you may not want to use long interleavers with short contest exchanges; the latency will need to be over 1 second to be effective.  You will need to add latency to the exchange time.  Selective fading happens quite often.  You can almost not avoid it with a Rayleigh path.

Anyhow, the solution is already at your fingertips, and the FCC has blessed it for years now.  It is called ASCII.  And 110 baud with 170 Hz shift is a breeze.

73
Chen, W7AY