9 May 2016

unid PSK-8 & MFSK-8 with LFM pulses (prob. Iranian source)

this transmission is composed of messages that are sent using standard MS188-110 Serial Tone waveform (the ones indicated as 3, 4 , ...) and other messages (as 9, 11, ...) that are transmitted using a waveform that most likely is a proprietary variant of 188-110. The transmission ends with a short op-chat (providing some clues about the source) and one interesting MFSK-8 modulated segment the closes the link and characterized by the presence of an LFM pulse waveform preamble.
Below the analysis of the message #11 only since it's the better recording among the not-standard 188-110 signals.

For what concerns the carrier, modulation and symbols rate, the signal shares the same parameters of an MS188-110 modem: PSK-8 on a single 1800 Hz carrier frequency and a constant 2400 symbols/sec output waveform (Fig. 1)

pic. 1
 The structure of the signal is indeed different: after a ~211ms sync preamble phase, the data phase consists of 51.25ms frames of alternating data and known symbols (Fig. 2). After 57 data frames a symbol sequence (most likely a subset of the initial preamble) is reinserted possibly to facilitate late acquisition, doppler shift removal, and sync adjustment as requested by 188-110 standard (Fig. 3).

pic.2 - sync preamble and data phases
pic. 3 - preamble re-insertions
The most peculiar aspect of this waveform is its data frame that counts 123 symbols or 369 bit (51.25ms, as indicated by the ACF function in Fig. 4). The data frame consists of 91 data symbols and a mini-probe of  32 symbols of known data (Fig. 5).
The length of the mini-probe, 32 symbols, is quite common and is largely used in 188-110 waveforms, including the appendix D and C. The oddity is the 91 symbols length of the data block. 

We will need additional recordings to indagate it.

pic. 4 - 51.25ms ACF
pic. 5 - frame structure
About the short op-caht in the final part of the transmission, a friend of mine suggests that the language may belong to the  Iranian group (Dari, Pushto, Kurdish) and possibly the protocol itself is developed there. That recording is available for who wants to indagate, simply email me.

The MFSK-8 segment the closes the link is shown in the zoomed FFT of Fig. 6.
pic. 6 - the ending MFSK-8 segment
It consists of three messages, each consisting of a Linear Frequency Modulation (LFM) pulse preamble followed by the same wavefom as MS188-141 (8 tones, manipulation speed of 125 baud and 250Hz step between carriers) but with different libraries since it is undecodable (Figs. 7,8).
pic. 7
pic. 8 MFSK-8 grid
update: comment sent by ANgazu
I agree with your analysis. Just an error, probably a typo one: preamble reinsertion is in every 42 frames and lasts for about 2 frames, so the superframe should be 40 data frames + reinserted preamble ( 2.160 s) or preamble + 40 data frames.
The preamble is much longer than annex C (about 110  ms in this case) and probe is 32 symbols instead of 31, so I agree with you that this looks like a new variant.

That's correct ANgazu, thanks !

No comments:

Post a Comment