A.3.1 32-Bit Floating Point or Integer Format
Least-significant 16 bits of floating point or integer numbers are placed at the smaller Modbus
address as specified in Open Modbus TCP Specification, Release 1.0. This is also known as
"little-endian" or "Modicon" word order. Floating point values are represented in standard IEEE
32-bit floating point format.
Despite the specification, some Modbus masters may expect "big-endian" word
order (most-significant word first). In such case, you must select "word-swapped"
floating point format in your Modbus master for the Modbus registers of the
device.
A "quiet NaN" value is returned for unavailable values. A quiet NaN is, for example,
7FC00000
hex
; however, the master should understand any NaN value.
A complete 32-bit floating point or integer value should be read and written in a
single Modbus transaction.
A.3.2 16-Bit Integer Format
Some 16-bit integer values in the data registers are scaled to include the necessary decimals.
The scaling factors for those values are shown in the register tables.
Table 36 16-bit Signed Integer Format Details
Value
Description
0000
hex
… 7FFE
hex
Value in range 0 … 32766
8002
hex
… FFFF
hex
Value in range -32766 … -1 (2’s complement)
8000
hex
Value is not available (quiet NaN)
Some values may exceed the signed 16-bit range even in normal operation. To
access such values, use the floating point registers instead.
Appendix A – Modbus Reference
47
Summary of Contents for HUMICAP HMP4
Page 60: ...www vaisala com...