The assembly of lag sets by the correlator backend is determined by the backend configuration. Each lag set belongs to a correlator output product defined by the backend configuration, and the binary lag set files produced by the backend are organized by output product. Within any such file, the lag sets are ordered by time, and lag sets appear in a sequence without any extra padding between adjacent lag sets.
Aside from the requirement that all lag sets in a file in the binary lag set format belong to the same output product, no other requirement is imposed on the files. In particular, a file may contain any number of lag sets, and the naming of files is left unspecified.
A schematic representation of a lag set in the binary lag set format is given in Table 1, “Binary lag set”.
Any frames of a lag set that were not received by the correlator backend (within some timeout interval) are, nevertheless, present in the lag set as recorded in the binary lag set format. All of the fields in the lag set associated with the missing lag frames are set to zero. Thus, the size of a lag set is fixed for a given output product, and all of the records in a sequence of lag sets in an output product have the same size.
Table 1. Binary lag set
| block | word offset | data word | |||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 28 | 24 | 20 | 16 | 12 | 8 | 4 | 0 | ||||||||||||||||||||||||||
Lag set headers | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 1 |
| 1 | BBID-Y | SBID-Y | SID-Y | BBID-X | SBID-X | SID-X | |||||||||||||||||||||||||||
| 2 | Num_Frames (= M + 1) | ||||||||||||||||||||||||||||||||
Lag frame 0 headers | 3 | Start_BlockY | NBlocks | nlags | ChipID | CCC# | CS-Valid | ||||||||||||||||||||||||||
| 4 | STATUS_BITS | FRAME_COUNT | RECIRC_BLK-Y | RECIRC_BLK-X | |||||||||||||||||||||||||||||
| 5 | TIMESTAMP-0 | ||||||||||||||||||||||||||||||||
| 6 | TIMESTAMP-1 | ||||||||||||||||||||||||||||||||
| 7 | DVCOUNT-Center | ||||||||||||||||||||||||||||||||
| 8 | DVCOUNT-Edge | ||||||||||||||||||||||||||||||||
| 9 | Board S/N | LTA Bin # | |||||||||||||||||||||||||||||||
| 10 | Checksum | ||||||||||||||||||||||||||||||||
Lag frame 1 headers | 11…18 | … | |||||||||||||||||||||||||||||||
| ⋮ | ⋮ | ⋮ | |||||||||||||||||||||||||||||||
Lag frame M headers | (3 + 8M) … (10 + 8M) | … | |||||||||||||||||||||||||||||||
Lags | 11 + 8M | Num_Lags (= N + 1) | |||||||||||||||||||||||||||||||
| 12 + 8M | Lag 0-In_phase | ||||||||||||||||||||||||||||||||
| 13 + 8M | Lag0-Quadrature | ||||||||||||||||||||||||||||||||
| 14 + 8M | Lag1-In_phase | ||||||||||||||||||||||||||||||||
| 15 + 8M | Lag1-Quadrature | ||||||||||||||||||||||||||||||||
| ⋮ | ⋮ | ||||||||||||||||||||||||||||||||
| 12 + 8M + 2N | LagN-In_phase | ||||||||||||||||||||||||||||||||
| 13 + 8M + 2N | LagN-Quadrature | ||||||||||||||||||||||||||||||||
Comments on some of the fields of the binary lag set format are provided below. Any field in the format that is not associated with the comments below is unchanged from the homonymous field in the LTA output data frame format. Note, however, that the endianness of the words in the binary lag set format is left unspecified (in contrast to the LTA data frame format).
Synchronization word; also useful for determining the endianness of the binary data words.
Number of frames comprising the lag set.
Boolean value representing the validity of the frame checksum. If the frame checksum is invalid (as determined by the correlator backend), the value of this field is zero; otherwise, the value is non-zero. This field exists only for convenience; the original LTA output frame could always be reconstructed, and the checksum validated by the reader of a lag set.
Number of complex lags in the lag set.
The lags are ordered as determined by the output product specification (given as part of the correlator backend configuration).