I recently had the chance to have a good copy of MFSK-n signal previously reported here as a MFSK 68-tone (34+34). Well, according to my new measurements, I tend to reconsider it as a parallel 64-tone (32+32) signal: this way the shift of 46.9 Hz among the tones is closer or almost the same of the apparent speed of 47 symbols/sec.
PSK-8 9000Bd bursts (2 symbol element periods length) are sent each 1 second, possibly acting as sync (Fig. 1): since the embedded PSK inserts, the signal occupies a bandwidth up to ~10KHz.
 |
| Fig. 1 |
It's worth noting that in the first half of the transfer a 3-tone symbols alphabet is used, while a 5-tone symbols alphabet (the most frequent in this wavefrom) is used in the second half, from time t1 (Fig. 2). In both the cases the speed of 45 symbols/sec remains constant (Fig. 3)
 |
| Fig. 2 |
 |
| Fig. 3 |
The measured 47 symbols/sec from SA likely results from the add of the two symbols due to the PSK-8 inserts.
The 64-tone solution can be easily seen by counting the tones over the -40dB threshold in the upper and lower part of the signal plotted in the FFT spectrum /Fig. 4).
 |
| Fig. 4 |
The number of the used tones remains constant, as well as the speed, for both the two alphabets (Figs 5,6) although the count of the grid in the second half of the signal (5-tones symbols) is more difficult.
 |
| Fig. 5 |
 |
| Fig. 6 |
Often you can see the CIS-3000 PSK-8 3000Bd modem before MFSK-64 as in Fig. 7: it's interesting to note that CIS-3000 and PSK-8 inserts have the same sub-carrier value (~5300 Hz) and it just matches the center of the MFSK-64 FFT spectrum (Fig. 8)
 |
| Fig. 7 |
 |
| Fig. 8 |
MFSK-64 is likely a Russian Intel parallel tones system as well as other MFSK and CIS-3000 could be employed as a sort of selcall.
https://yadi.sk/d/WMnwnpN034qLZV
