SkyOFDM, still on air

Since some days I'm following transmissions on 14693.0 KHz/USB consisting of exchanges of messages between two nodes, as the different fading patterns in Figure 1 suggest.

Fig. 1

Messages are sent using OFDM modulation occupyng a 2400 Hz bandwidth and consisting of 28 tones with a frequency spacing of ~86 Hz, each tone is modulated using PSK2 at the symbol rate of 62.6 Bd (Fig. 2). The same results are obtained/verified by analyzing a single channel as shown in Fig. 3 (lowest tone).

Fig. 2 - OFDM parameters

Fig. 3 - single tone (yhe lowest) analysis

The parameters resulting from the analysis are very similar to those of the Skysweep Technologies proprietary "SkyOFDM" waveforms family (Table I). Quoting the SkySweeper Reference Manual #82.2 General Description: "SkyOFDM is a state of art high speed modem based on the OFDM and turbo coding technologies. It offers several baud rates (300 -9600 bps) and two different interleaving options (short and long). Also there are two bandwidth options: 2.0 (OFDM-22) and 2.4 kHz (OFDM-28)".

Table I

Note the different number, position and duration of the header tones compared to the values ​​of the "original" waveform: this is probably an improved version of the previous SkyOFDM waveforms (Figure 4). 

Fig. 4

The signal has an ACF value of approximately 957.8 ms which identifies a super-frame composed of 11 frames, the latter with an ACF value of approximately 79.8 ms (Figure 5).

Fig. 5 - ACF values

Direction Finding tests using TDoA algorithm (Figure 6) indicate an area north of Helsinki as the site of the transmitter (or rather, the radiating antenna): this makes sense because, acccording to some DXers, SkyOFDM waveforms were/are used by Finnish MFA and SkySweep Technologies was a Finnish high tech company. By the way, although there are still many references in the web to SkySweep, their official website is no longer online since SkySweeper software was discontinued on June 1st 2009.

Fig. 6 - Direction Finding

 https://disk.yandex.com/d/A9UHdyMAXeOUlA

CIS FTM-4 transitions

CIS FTM-4 (FTM stands for Frequency-Time Matrix) is an unknown Russian "domestic" system, also known as CIS 4FSK 150 Bd, using MFSK4 150Bd/4000Hz modulation.
The recorded transmission consists of two alternating FTM-4 sets (L = lower set, H = higher set) with a total bandwidth occupation slightly over 25 KHz, the separation between the two sets (i.e. H1-L4) is 1 KHz. It's worth noting that when the L set is transmitted, the signal on the H1 frequency (i.e. the lowest of the H set) is continuous; vice versa, when the H set is transmitted the signal on the L1 frequency (i.e. the lowest of the L set) is continuous (Figure 1). Also note that the two sets are transmitted simultaneously forming a sort of MFSK-8 "construct" for a period of about 3 seconds.

Fig. 1 - the two FTM-4 "sets"

As for FMT-4 specifications, both the MFSK4 sets are modulated at the rate of 150 Baud with a spacing of 4000 Hz (Figure 2).

Fig. 2 - CIS FTM-4 main parameters

The L set has an ACF value of ~718 ms which corresponds to a repeated sequence of 216 bits length (assuming that MFSK4 uses 2 binary digits (dibit) per modulation symbol (0-3), the ACF value of 718 ms @150 symbols/sec corresponds to a period of 108 dibit symbols or 216 bits). I could not find a specific generator polynomial for that sequences.

Fig. 3

The study of the ACF value of the set H can be done on three different intervals: main, intermediate and unitary (Figs. 4, 4b). 

Fig. 4

 

Fig. 4b

 

The main ACF measures about 1920 ms which corresponds to a period of 288 symbols or 576 bits (Figure 5).

 

Fig. 5

The intermediate ACF measures about 480 ms which corresponds to a period of 72 symbols or 144 bits (Figure 6).

 

Fig. 6

The unitary ACF has a value of ~80 ms which corresponds to a period of 12 symbols or 24 bits (Figure 7).

Fig. 7

In this case (H set) it is possible to see that the demodulated bitstream consists of a repeated scheme formed of 4 patterns, as in Figure 8:

(MSB first)
P1: 0x21BE41
P2: 0xB1BE41
P3: 0xB8D727
P4: 0x28D727 

Fig. 8

 

N.B. all the the "designations" I used here are only mine; the first pattern P1 is chosen just for convenient reference, choosing a different initial pattern the "logic" does not change.

A possible interpretation (just a guess) is that the L set is transmitted as "idle" or alignment sequence for the receiving modem and the transition to the H set occurs when sending data, even if, as in this case, the data are repetitive sequences. In this regard, remaining in the realm of hypotheses, the P1-P4 patterns could be telemetry data or tele-commands.

The recording was made using a remote Airspy HF+ Located in Haapavesi, Finland (belonging to the Airspy server network) [1].

https://disk.yandex.com/d/Flr5-Ops8dRBJQ

[1] sdr://178.55.138.222:5000