Campbell Scientific RTDAQ Software - Manual (Page 361)

Appendix A. Campbell Scientific File Formats
Header line five describes the processing performed in the datalogger to
produce the value for each field in the record; for example, sample, average,
min, max, etc. If there is no known processing for a field, that field will be
assigned an empty processing string. There will be one value on this line for
each field name given on header line three.
Header line six defines the data types for each field in the record and supports
the following values: IEEE4, FP2, ULONG, LONG, SecNano, and ASCII
(len).
TOB2 frame headers are eight bytes long and hold the timestamp for the first
record in the frame. TOB3 frame headers are twelve bytes long and contain
the same timestamp information but also add a four-byte unsigned integer that
represents the beginning record number for that frame.
The frame data begins immediately following the frame header and consists of
zero or more data records. Each record contains one data point for each of the
field names identified in header line three. The data type and implied size of
these data points are identified by the data types list given by header line six.
The frame footer makes up the last four bytes of the frame.
A.3 Binary Data Value Types
When data is written in datalogger memory or in binary data files each value
must be assigned a particular data type. These data types describe the format
of the data.
A.3.1 FP2 (2 Byte Low Resolution Format)
A two-byte floating-point number format created by Campbell Scientific, Inc.
and used to store low-resolution values. Basically, this format consists of a
single sign bit, a two-bit negative decimal exponent, and a 13-bit mantissa.
For more detail, refer to the CR1000 manual or the CRBasic Help.
A.3.2 FP4 (4 Byte High Resolution Format)
A four-byte floating-point number format created by Campbell Scientific, Inc.
and used for input location values as well as high-resolution final storage
values. This format consists of a single sign bit, a seven-bit base-two
exponent, and a 24-bit mantissa.
A.3.3 IEEE4
A standard four-byte floating-point number format used for certain values
within a record. This format consists of a single sign bit, an eight-bit binary
exponent, and a 23-bit mantissa.
A-5