On Jan 27, 2010, at 1/27 4:32 AM, Bill, W6WRT wrote:
> As I recall, after the High Speed Sprints of the 1990's, the general
> consensus was there were too many errors caused by propagation
> vagaries. 45 baud simply prints better in most cases.
That is correct.
Changing from 45.45 baud to 100 baud will (a) widen the required
bandwidth (not by a factor of two since we don't change the 170 Hz
shift) and (b) reduce the bit period by a factor of more than 2 from
22 ms to 10 ms. Taken together, 100 baud will have a significant
(something over 4 dB, if my back of envelope calculation is correct)
loss in SNR when all else (power, etc) is equal, with a corresponding
rise in character error rates vs. 45.45 RTTY.
To see why (a) and (b) affects error rates, take a look at this
http://sss-mag.com/ebn0.html
The note from Phil, KA9Q, about a third of the way down in that page
relates Eb/No to SNR (or CNR, the carrier to noise ratio), which is
easier to relate to SNR plots such as the plots for RTTY at VE3NEA's
web site. After the SNR conversion (using Eb/No and then Phil's
conversion back and forth from and to SNR) for 100 baud RTTY, you can
see what the amount of SNR loss does when compared to 45.45 baud
RTTY. (Phil was the one who wrote the TCP/IP stack for Amateur bands,
if you can remember that far back :-). You can then get a sense from
VE3NEA's plot how 100 baud RTTY performs relative to 45.45 baud for
stable non-multipath propagation (AWGN), flat fading, selective fading
and flutter. Remember that Alex VE3NEA had use 3 kc noise bandwidth
as his calibration point -- so Phil's equation would need to use a Bw
value of 3000 if you are going to use it with Alex' plots.
73
Chen, W7AY
_______________________________________________
RTTY mailing list
RTTY@contesting.com
http://lists.contesting.com/mailman/listinfo/rtty
|