[RTTY] Jitter

[Kok Chen]

> On Feb 26, 2018, at 12:41 PM, Hank Garretson <w6sx at arrl.net> wrote:
> 
> Question: Does my jitter increase my transmit bandwidth?

Not necessarily much wider (it won't be narrower; think "entropy" :-).    But it will sound "noisy" to ears that are used to clean RTTY.

It is for a different reason that you want jitter to be low.

> Or put another way: If I switch to TinyFSK, will my transmit bandwidth be
> reduced?

You really need good timing... and that is what TinyFSK gives you.

The way to visualize it is this...

You start by generation rectangular wave pulses that correspond to whether you want to transmit Mark or Space.

By the time the decoder sees the pulses that you have sent, they have turned into triangular (if you are using Matched Filters) or sinusoidal waves (if you are using optimal raised cosines).

You no longer have wide rectangular pulses at the decoder.  They peak at some "mid-bit" locations.  If you jitter, the decoder is no longer sampling the true "mid-bit" and you therefore lose SNR.

You can see the triangular waveforms for Matched filters in the last three figures of the web page I had previously mentioned:

http://www.w7ay.net/site/Technical/RTTY%20Sidebands/sidebands.html

For the optimal raised cosine filters (2Tone uses them), see the last figure here:

http://w7ay.net/site/Technical/EqualizedRaisedCosine/index.html

> If yes, by how much?

If the signal starts with a very good SNR (your neighbor's RTTY signal), then your peak-to-peak jitter can be 11 millisecond before you see degradation for a 45.45 baud RTTY signal.

If you are using 75 baud, the peak-to-peak jitter of 6.5 milliseconds will cause errors.  This is why most people consider bit-banged FSK from a computer to be unusable for 75 baud RTTY.  (TinyFSK is also a computer, but every processor clock cycle is accounted for.)

Remember that multi-path will already cause the mid-bit to move around randomly.  Jitter simply exacerbates it.

73
Chen, W7AY