Автор Тема: WSPR-15 in Dreamers Band  (Прочитано 90 раз)

0 Пользователей и 1 Гость просматривают эту тему.

Оффлайн rn3aus

  • Ветеран
  • *****
  • Сообщений: 2613
    • Просмотр профиля
WSPR-15 in Dreamers Band
« : 10 Март 2020, 20:55:13 »
Маркус DF6NM провел интересный эксперимент, используя способ декодирования WSPR, описание которого находится здесь: http://136.su/index.php/topic,156.msg27178.html#msg27178

Из рассылки:

On 8.27 kHz VLF, aka Dreamer's band, signal levels are usually too weak for anything but long carrier detections or dedicated EbNaut transmissions. But we have seen significant improvements in receive capabilities from three axis receivers, like Paul Nicholson's in Todmorden, DK7FC in Heidelberg, and DL0AO in Amberg. So I wondered whether WSPR-15 might also play a role there.
Last weekend we installed a wsprd decoder at DL0AO, driven by a dedicated 3.8 mHz SpecLab instance. After fixing a couple of stupid errors (like forgetting to bypass the FFT-filter block, and converting 8270 to 1500 instead of 1610 Hz), my weak signal was quite visible in the spectrogram, but was still not decoded. I wondered if something was wrong with the transmitted sequence?
In the end, I discovered that the WSPR decoder cannot cope with messages which cointain a negative dBm (<1 mW) power entry. This was possible in early versions of the software, and the WSPR-X and WSJT-X drop-down menues still go down that far. But apparently, one of the power-encoding bits has later been reallocated for new compound-type messages. So my true-power "DF6NM -20 dBm" messages were just quietly dropped!

In the mean time, several hours had gone by and the afternoon QRN had come up. So the next attempt had to wait until Monday morning. And that was finally successful, with better than expected SNR (albeit overstated power):
3 spots:
    Timestamp      Call   MHz    SNR Drift  Grid   Pwr   Reporter   RGrid   km   az
 2020-03-09 05:30   DF6NM   0.008270   -33   0   JN59nj   0.001   DL0AO-VLF   JN59vk   48   84
 2020-03-09 05:15   DF6NM   0.008270   -32   0   JN59nj   0.5   DL0AO-VLF   JN59vk   48   84
 2020-03-09 05:00   DF6NM   0.008270   -33   0   JN59nj   0.5   DL0AO-VLF   JN59vk   48   84
Can this be compared to EbNaut? On one hand, WSPR requires around 7 dB Eb/N0 for its 50 bit payload, compared to ~ 0 dB for EbNaut, which could be five times faster at same power. On the other hand, requirements on frequency/phase stability and symbol timing are much more relaxed, so you can get away with a freerunning samplerate and manually initiated start of transmission.
The DL0AO decoder continues to lurk for WSPR-15 signals in the 8260 to 8285 Hz band. The next step may come soon: Stefan has been preparing to send WSPR-15 from his DK7FC/p earth antenna, and chances are that this can be decoded both in Heidelberg and Amberg.
All the best