By Murray Greenman, ZL1BPU
Domino is the name given by the developers to a family of IFK coded coherent phase single tone MFSK keyed modes, using sequential tone-pairs in two alternate fields arranged as orthogonal but interleaved tone sets.
The name ZL2AFP domino is that given to this first software realization of these modes.
The main aim of domino is to provide a simple-to-use, but slick-to-operate, digital chat mode for HF, especially the lower bands (160/80/60/40/30m) where multipath is such a problem. It was designed for beginners with limited skills, and perhaps equipment with limited performance, especially in the drift and offset department.
MFSK offers very good immunity to interference and ionospheric effects, and is also very sensitive. However, it has several disadvantages, which remain problems even with good FEC. These disadvantages make the MFSK modes unattractive to beginners and those wanting fast exchanges of information. The problems are:
- An extreme tuning accuracy requirement, for example 4 Hz on MFSK16 and less on THROB and MFSK8
- Low drift tolerance, typically ±4 Hz per minute
- Slow sync lock at start of transmission (seconds)
- Latencies introduced by the FEC process that reduce "slickness" (6 sec in MFSK16)
- Difficulty identifying a weak signal for tuning purposes, as the lowest or center tone needs to be found exactly.
This new family of domino modes attempts to address these disadvantages with varying levels of success, and some trade-offs. The specific measures employed (so far!) in domino are:
- Incremental Frequency Keying
IFK provides independence of tuning and tolerance of drift. The cost is increased errors (one error always means two errors). There are numerous IFK related error mechanisms that still require better understanding. The IFK technique was invented by Steve Olney, VK2ZTO, and has been successfully employed on LF by Alberto, I2PHD, in the JASON program. This is its first ever deployment in an HF chat mode.
- Alternating Tone Sets
Two independent tone sets, used alternately, to provide very fast symbol synchronism. The alternation is very quickly detected as an odd-even component in the receiver FFT, and is much faster than the impulse response technique used in MFSK16. Sync lock-up time is well under one second. The cost is a waste of useful bandwidth.
- Orthogonal Tone Sets
Independent tone sets to provide very good ISI performance. The improvement noticed so far is mostly in the time domain. The tone sets are arranged so that ionospheric modulation products cannot easily spill from one possible tone position to another, or from one symbol to the next. This is done by double-spacing the tones and interleaving the tone sets. A paper by Tim Giles triggered this idea. There will be better ways to achieve this, perhaps using a tone pattern restriction rule as suggested to the authors by Nino IZ8BLY, rather than two separate tone sets.
- Tone Set Interleave
Interleaving the tone sets obviously considerably reduces the necessary bandwidth, but the cost is confusion at the receiver, where due to receiver tuning errors, the FFT cannot know from which tone set a particular tone comes. We get round this with a fudge factor, sending a unique non-printing character that is detected readily when the tone order is wrong.
- Character-based System
Two symbols are used to send each character. The domino design is character based, not a binary system or bit-based system such as PSK31 or MFSK16. Each pair of tones means a specific character. The Coquelet designers assert that 2-tone MFSK gives lower BER than other multi-symbol systems, and slightly better than binary systems.
- Low Symbol Rate
A low baud rate is used to defeat the effects of multi-path reception. The slowest mode, domino8, will easily handle 100ms of multi-path, making it useful under the absolute worst of NVIS conditions, as experience has shown.
- Reduced Character Set
A small character set is used (63 characters), to improve efficiency at low data rates. Even at 11 baud, the typing speed is 44 WPM.
- No FEC Coding
No FEC is used as the mode is reasonably robust without it.
It is this combination of techniques that give this mode its unique character. The performance that results, while not stunning, and not without other problems, is never-the-less good.
IFK provides high tolerance of receiver drift and tuning offset. Unlike MFSK16, which requires a tuning accuracy of 4 Hz and a drift rate of less than 4 Hz per minute, thanks to IFK, domino will handle an offset of 200 Hz or more and drift of 200 Hz/minute, provided of course that the signal remains within the operating passband.
Because of the double-spaced and interleaved tones, and the low data rate, domino can handle about 100ms of multi-path error (by comparison RTTY copes with less than 5ms). Obviously, domino16 is more affected than the slower domino11 and the slowest domino8.
With a restricted character set and no FEC, a data rate of 44 WPM is achieved at a symbol rate of only 11.025 baud. This compensates for the reduced data bandwidth due to interleaved alternated symbol pairs (16 tones are used for only three data bits instead of four). The mode is sufficiently robust that FEC is not usually needed.
The insistent "falling" sound to the tones when idle or between words can identify IFK11, and tuning is easy.
Domino encompasses three modes, identical except for baud rate, speed and bandwidth. These are achieved primarily by changing the sound-card sampling rate to 8000, 11025 or 16000 samples/sec. It is realized that not all sound cards will support all three sampling rates, but this technique reduces receiver overhead, allowing the software to operate on quite slow computers.
Multi-tone frequency shift keying, single tone coherent phase (CPMFSK), using 16 tones spaced 1T, but used as orthogonal fields of 8 tones spaced 2T, interleaved at a spacing of 1T.
Each 3-bit data octet is multiplied by two (shifted left), and summed with a symbol synchronous square wave at half the baud rate with values 0 and 1. This results in two orthogonal tone fields with 2T spacing in-field, and each tone is at least 1T and one symbol period spaced from any other tone.
Characters are incrementally coded in two successive 3-bit data octets representing a proprietary 6-bit character set. Each pair of octets is subtracted (using signed addition) from the corresponding coded octet (i.e. the previously transmitted octets) in the previous character, and the signed result of each octet is logically ANDed with 7 to keep it within range. (At the receiver each octet pair is added to the last decoded corresponding previous octet and ANDed with 7). The high octet is transmitted first. Thus, the increment is signed and 6 bit, based on subtracting the new uncoded character from the previous transmitted character.
Tone set order (character sync) is recovered at the receiver from data context. This is possible because frequently used characters have unused equivalents when viewed with octets swapped. A special idle character is sent every 16 characters, making the reversal more obvious.
Using the ITU definition, the necessary bandwidths for the modes domino8, domino11 and domino16, with symbol rates of 7.8125, 11.025 and 15.625 baud, are 158, 223 and 316Hz respectively. (45 baud RTTY rates 453Hz)
The formula used for MFSK mode bandwidth calculation is BW = B + msk where B = baud rate, m is the number of tones, and s is the tone spacing. k is a factor related to keying technique, and k = 1.2 for coherent phase keying.
ITU Emission Designator
The mode is single channel subcarrier modulated SSB transmitted automatically received data, so rates as J2B. Hence, the modes are classified as 158HJ2B, 223HJ2B and 316HJ2B. (By the way, the last of these is identical in designator and bandwidth to MFSK16, while RTTY rates 450HJ2B). Emission Designators are outlined in FCC Part 47 Rules, para. 2.201.