IRIG 106 Chapter 4: what the PCM telemetry standard actually requires

Where Chapter 4 sits

IRIG 106, published by the Range Commanders Council, is the reference set of telemetry standards for flight testing. Chapter 2 covers the radio link, Chapter 4 the PCM stream that rides on it, Chapter 9 the TMATS setup record that describes the stream to the ground, Chapter 10 the recorders. When a ground station "supports IRIG 106 PCM", Chapter 4 is what it supports.

Chapter 4 defines the structure of the stream, not the meaning of its words. Which parameter sits in which word, with which scaling, is described in TMATS — the setup record shared by the airborne encoder and the ground.

Class I and Class II

The chapter divides PCM formats into two classes. Class I is the simple, conservative set every range can receive. Class II covers the more complex techniques, and using any of them requires the concurrence of the range involved. Staying in Class I means a format can be received anywhere without discussion; going to Class II is a decision to make with the range, early.

The bit stream

• Six codes are allowed, one per stream: NRZ-L, NRZ-M, NRZ-S and the three bi-phase codes, Bi-L, Bi-M and Bi-S. Randomised NRZ-L is NRZ-L passed through the standard's 15-bit randomiser.
• The stream must be continuous and carry enough transitions for the ground to keep bit synchronisation, which is what randomisers and bi-phase codes are for.
• The minimum bit rate is 10 bit/s; above 10 Mbit/s is Class II, and the radio and recording limits set elsewhere in IRIG 106 bound the maximum.
• While data are being sent, the bit rate must stay within 0.1 % of its nominal value, and jitter within 0.1 bit of the expected transition.
• Clock phase is defined as well — 0°, 90°, 180°, 270° or double data rate — so that encoder and recorder agree on when a bit is valid.

Words and frames

• Words of different lengths may share a minor frame, as long as each word position keeps the same length in every frame.
• The minor frame opens with a synchronisation pattern of 16 to 33 bits, always counted as word zero; the chapter's appendix lists recommended patterns.
• A frame counter in every minor frame is recommended. With subcommutation, the subframe counter (SFID) plays that role: it sits in a fixed position, ahead of the subcommutated words.
• A major frame contains every word of the format at least once, and holds at most 256 minor frames. Subcommutation depth runs from 2 up to the number of minor frames in the major frame.
• In Class I, supercommutated samples must be evenly spaced.
• Numbering is fixed so that everyone means the same thing: bit 1 is the most significant bit of a word, word 1 is the first after the sync pattern, and minor frame 1 opens the major frame.

Time in the stream

Chapter 4 standardises time words, so the ground does not have to guess how an encoder wrote its clock. Three 16-bit words carry it: a high-order word, a low-order word with a 10 ms resolution, and a microsecond word that counts from 0 to 9,999 before rolling over. High and low order may be binary or BCD; in binary, one unit of the high-order word is 655.36 s, in BCD it is one minute, with part of the day of year alongside.

The recommended order is high, low, then microsecond, ahead of the first data word of the minor frame. For other word sizes, the time words are packed into contiguous PCM words and padded with zeros. The microsecond word alone can time-tag individual data words, provided the other two come often enough to resolve the ambiguity.

Class II tools worth knowing

• CRC: a 16-bit (ANSI or CCITT) or 32-bit CRC at the end of each minor frame, computed in transmission order with no inversion and no unusual starting value, so that any ground software can check it without vendor-specific options.
• Format changes: up to 16 formats, switched on minor frame boundaries and identified by a frame format identifier present in every minor frame, built so that a single bit error cannot turn one valid identifier into another. They disrupt processing; fixed formats are preferred.
• Asynchronous embedded formats: a second PCM stream carried in dedicated word positions of the host frame.
• Asynchronous data merge: serial data such as RS-232 inserted in fixed word positions, each with a stale-data bit and an overflow bit, so the ground knows whether a value is new and whether something was lost.
• Tagged data: blocks of words with identifiers. For MIL-STD-1553 and ARINC 429, the preferred method remains to restructure the bus data into Class I words; otherwise the bus formats of Chapter 8 apply.
• Packets: Chapter 7 carries Chapter 10 packets, Ethernet frames or video inside the minor frame, alongside the classic words.

Checking a format before the first flight

1. List every Class II feature the format uses, and get the range's agreement for each one.
2. Check the bit rate against the transmitter, the receiver bandwidth and the recorder, not only against the 10 Mbit/s line.
3. Check that the TMATS file matches the encoder configuration word for word: sync pattern, word lengths, frame lengths, subframe counter position.
4. Decommutate a recorded test stream on the actual ground equipment, and confirm frame lock, subframe lock and time words.
5. Feed a known value on a few channels and check that it arrives in the expected word, at the expected bits and scale.