[aprssig] 6 meter APRS

Stephen H. Smith wa8lmf2 at aol.com
Wed Aug 26 15:02:16 EDT 2009


Chris Moulding wrote:
> Hi,
>
> One project I'm working on in whatever spare time I have is a PSK-125 
> TNC add-on for UI-View.
>
> To test this I've already written a macro in Fldigi to simulate an 
> APRS message.
>
> I've just spent a few minutes running the message through various digi 
> modes and CW speeds to
> measure the time taken. See the results below:
>
> The test message is:
>
> G4HYG>APRS,GATE,WIDE2-2:=5332.76N/00225.91W&Chris, testing PSK modes 
> for APRS
>
> Time taken to send the message:
>
> PSK-31................23.66 s
>
> PSK-63................12.30 s
>
> PSK-125................6.91 s
>
> PSK-250................3.87 s
>
> RTTY (45 bd)......18.50 s   (note that > is not in the character set)
>
> OLIVIA (8-250)....59.48 s
>
> CW (18 wpm).......64.43 s
>
> CW (50 wpm).......23.61 s
>
> CW (100 wpm).....11.66 s
>

The biggest problem with data transmission on HF is multipath 
propagation causing the same bits to arrive at the receiver over several 
different paths (of different length and time delay) simultaneously.  
People liken HF propagation as bouncing signals off the "mirror" of the 
ionosphere, but it is more like bouncing a searchlight off a fog bank. 
There is no coherent single surface doing the reflecting. This causes 
sequential bits to overlap in time and smear each other if they are sent 
too fast.

The key to reliable HF data transmission is to use relatively low symbol 
rates (i.e. duration of each symbol substantially longer than the 
potential spread of propagation times of 10's of milliseconds.)   This 
is why hand-send CW and PSK31 are so effective under conditions of 
less-than-ideal propagation. Their individual symbols are LONG! 

To up the data rate, one has to increase the information content (i.e. 
complexity) of each symbol, not increase the symbol RATE.  This means 
more phases, more amplitudes or more simultaneous carriers but not 
FASTER symbol rates.  

Have you considered QPSK rather than the more common BPSK?  It is 
supported in just about every PSK program, and can double the data 
throughput without increasing the symbol rate.   




More information about the aprssig mailing list