LF: Op32 correlation detection: RN3AGC, UA4WPF, UW8SM

[Markus Vester <markusvester@aol.com>]

Last night Jan-21/22 I tried again to detect Opera signals using the "deep search" correlatiopn technique.

 

Attached is a spectrogram covering the 137.5 to 137.6 kHz segment in 0.477 mHz FFT resolution. At first, the strong WSPR signal from DK7FC on 137.607 completely defeated the noise blanker. This could later be mitigated by inserting a steep bandfilter in SpecLab. The spectrum shows a large number of peaks, most of them from active Loran chains, eg. 6.25 Hz multiples from GRI 8000.

 

The FFT results were exported in ten minute intervals and processed offline by my experimental Opera detection software, which searches for non-Loran peaks and correlates the spectrum in their vicinity against several known Opera signatures. Phase-coherent and incoherent searches were done in two separate runs. This produced the following hits:

 

coherent:

 

 frequency       peak     focused    Opera       end time   corr.  callsign 
137536.953Hz 10.0dB 20.7dBaf -46.9dBOp  20:42:13 15.5dB UA4WPF
137536.953Hz  7.4dB 17.4dBaf -50.2dBOp  21:18:12 16.0dB UA4WPF
137556.953Hz  6.2dB 18.1dBaf -49.5dBOp  21:26:11 14.1dB RN3AGC
137556.938Hz 14.2dB 26.2dBaf -41.4dBOp  22:13:29 17.5dB RN3AGC
137556.953Hz 15.4dB 27.3dBaf -40.3dBOp  23:09:31 20.1dB RN3AGC

 

incoherent:

 

 frequency        peak   focused    Opera       end time   corr.  callsign 
137556.938Hz  5.6dB 28.6dBaf -39.0dBOp  22:13:29 17.8dB RN3AGC
137556.938Hz  6.5dB 32.7dBaf -34.9dBOp  23:09:31 20.1dB RN3AGC
137562.547Hz  4.6dB 31.5dBaf -36.1dBOp  00:08:18 15.8dB UW8SM
137562.547Hz  6.5dB 27.3dBaf -40.4dBOp  02:31:47 14.4dB UW8SM
137562.531Hz  5.5dB 32.7dBaf -34.9dBOp  03:19:38 14.6dB UW8SM
137562.563Hz  5.3dB 31.7dBaf -36.0dBOp  04:07:26 19.7dB UW8SM
137562.547Hz  6.1dB 33.1dBaf -34.5dBOp  05:43:08 19.6dB UW8SM

 

UA4WPF and RN3AGC seem to be phase continuous whereas UW8SM could only be detected incoherent mode. Note that in the current state the SNR calibration still has large uncertainties, but the coherent mode can clearly see much deeper.

 

Unfortunately UW8SM seems to be sitting right on top of a Russian Loran line on 137562.5. A slight shift in frequency would (eg. 0.5 Hz) probably provide better coverage for him.

 

Best 73,
Markus (DF6NM)

-----Ursprüngliche Mitteilung-----
Von: Markus Vester <markusvester@aol.com>
An: rsgb_lf_group <rsgb_lf_group@blacksheep.org>
Verschickt: Sa, 8 Dez 2012 7:03 pm
Betreff: R7NT Op32 correlation detection

Alex, LF,

 

the Op32 signal from R7NT was well visible in the 21 mHz TA window. I made two recordings of a 2 Hz band around the anounced frequency, one from 22:08 to 08:11 UT and another one in daylight from 12:06 to 15:25. Then I ran both through my Opera correlation detector, see attached result.

 

The first two lines are a list of several known active Opera stations. Each of these was tried in the cross correlation procedure. The lines below that are the successful hits. Going from left to right, the printed output contains:

 

 time window for correlation (decimal UT hours before / after midnight last night),

 frequency of carrier peak,

 uncalibrated "raw" dB of detected carrier,

 carrier SNR after a "carrier autofocus" procedure, which removes small amouts of chirp and optimally weighs the amplitude for QSB,

 SNR on the Opera scale (ie. carrier in 2.5 kHz minus 4 dB),

 either c or i for a coherent or incoherent demodulation,

 UT hour of the winning correlation peak (corresponds to the start of the transmission),

 SNR of correlation peak in time domain,

 callsign of the winner of the correlation procedure.

 

Now running a recording on 137567 to look for UA0AET tonight.

 

Best 73,

Markus (DF6NM)

 

 

df6nm_Op32_130121-22.png
Description: PNG image