Stefan wrote:

> f = 4470.100000 Hz
> Start time: 18.JAN.2018  16:00:00 UTC
> Symbol period: 60 s
> Characters: 2
> CRC bits: 16
> Coding 16K23A (NEW)
> Duration: 13h20min [hh:mm]
> Antenna current: 300 mA

Todmorden: Eb/N0 +8.6 dB, S/N -27.7 dB/1Hz, phase 70.3 *Decoded*

Bielefeld: Eb/N0 -2.0 dB, S/N -38.3 dB/1Hz, phase -45.2;

Cumiana: Eb/N0 +8.9 dB, S/N -27.4 dB/1Hz, phase -116.3 *Decoded*

Warsaw: Eb/N0 -4.4 dB. Many interruption to the uplink and the
uplink failed completely at 22:21.   Even so, -4.4 dB.  It would
probably have decoded if the uplink stream had continued.

Bielefeld was +13 dB the previous night. What happened there to
lose about 18 dB?

Jim wrote:

> FFT -> excise hum lines -> IFFT,  before blanking?

That's how vtfilter works when given a -a option.  It works well
to remove harmonics, but not so well to remove modulation sidebands
produced by heavy electrical machines in Todmorden's industry.

--
Paul Nicholson
--