3 August 2015

SERDOLIK OFDM 35-tone DQPSK


This signal, heard several times and just today on 14390.5 USB starting from 0810z, is one of the many waveforms of the so-called SERDOLIK, a complex system used by the Russian Intel and Diplo services. Despite an initial annoyng CIS 36-50 interference (and the heavy selective fading), there are clearly visible the MFSK-34 (Crowd 36) parts and the OFDM one in the middle: the MFSK-34 partes are characterized by 40 Baud speed and 40 Hz channel separation, as shown here http://signals.radioscanner.ru/base/signal36/

The interesting part of this signal is the one related to the OFDM: another variant of the Serdolik OFDM 35-tone waveform shown here: http://signals.radioscanner.ru/base/signal123/ although the baudrate is the same (40Bd) as well as the used modulation (PSK-4).
On there, the OFDM part is formed w/out the cyclic prefix (ie channel separation = Br = 40Hz) and has 34 channels (the second probably used as pilot). On here, the OFDM part exhibits same baudrate (40 Baud) but 35 channels (the lower probably acts as pilot) with a 40 samples lenght cyclic prefix, channel separation = 50 Hz and k (correlation triangle) =1/4 (Pic. 1); the used modulation DQPSK


Pic. 1 OFDM params and modulation in SERDOLIK 35 modem PSK-4

OFDM parameteres and channels modulation  are the same of the 61 tones modem as reported here Serdolik 61.
First channel from bottom has no modulation and according to my measurements this lower channel is plus 6dB compared to the others (Pic. 2) just like a pilot tone does.

Pic. 2
Interesting is the behavior of the 3rd channel that looks like an "hole": looking deeper, it's a BPSK at 40Bd speed and KarapuZ thinks it acts as an additional sync since it carries only binary strings of zeroes and ones (Pic. 3 and 4), thanks for the tips.

Pic. 3-a

Pic. 3-b

Pic. 4

Since the different gaps between MFSK and OFDM channels the two signals are more likely not formed at the same stage, it's my guess that the initial segment of the OFDM part just acts as syncronization  (Pic. 5)

Pic. 5