Paul, Markus,

OK:, new constellation:

Code:Rate 1/18 K = 25
Characters: 25
10sec/symb
1520 signal bits
duration: 4h14sec

TX time: 11:00 to 16:00, day: when the WX will be fine.

Uwe/dj8wx

Uwe wrote:

> I would prefer Rate 1/8K = 25 out of your convolution code
> message generator, i.e. 560bits in 4h40min.

Your last transmission I think was on 2014-11-07 and I just
examined the recording: You were S/N 14.7dB in 278uHz between
12:00 and 14:00. Signal is strong enough to send data quite
fast. You could use 14 second symbols to send 10 characters
(800 signal bits in 03:06:40) for a virtually certain decode.

A more challenging test with 11 second symbols would still
have 95% chance of correct decode.

So far the best radio amateur message transmission at VLF was
by Markus DF6NM on 2014-05-18 when he sent 14 characters across
1028km. If we use product of distance and characters as a metric
then that's a score of 14392 kmchars.

At 762km you would have to send 19 or more characters to beat
that. Daytime should allow a 4 hour message without trouble.
Rate 1/8, 11 second symbols, 21 characters (1328 bits in 04:03:28).
Avoid the terminators. Send around midday +/- 2 hours.

On 2014-12-14 I wrote:

> http://abelian.org/vlf/tmp/141219c.png

> From 00:00 to 08:00 it averaged 12.6dB S/N in 278uHz, 0.14fT.
> That's enough signal to send 27 characters at a rate of 22
> information bits per hour.

I gave this a test using a recording of W4DEX carrier from
that night. Modulating the recording with a BPSK message 27
characters long, rate 1/16 K=25 with a symbol period of only 7.5
seconds and it decoded correctly with a small margin to spare.
9 second symbols decoded with plenty of margin.

--
Paul Nicholson
http://abelian.org/
--