Siemens MD-12HD, Руководство пользователя, Страница: 61 / 74

Siemens Industry, inc. 60
Floating Point Registry Entry
The MD-xxHD meter uses 32-bit IEEE 754 formatted floating point numbers
for reporting results and storing scalable user register values such as CT
range, CT and PT scaling factors, etc. Because these registers require two
16-bit Modbus addresses to convey, these registers must be accessed as
multiple registers or accessed sequentially without interruption. The reason
for preventing floating point registers from being updated as single 16-bit
registers is that interim values (when the number is half entered) represent
valid but unknown numeric values! Requiring both the MSW and LSW
registers to be written sequentially prevents meter data from having unknown
and potentially very large scaling factors applied to measurement data
between register writes!
SELECTING THE DATA TYPE
It is likely that the RTU program has built-in support for multiple data types
including floating point. The MD-xxHD meter data is stored as MSW, LSW
which may take some trial and error to identify in the RTU setup. Float ABCD
is an example of the how this RTU identifies the matching byte order.
It is anticipated that command-line programmers or script writers may prefer to enter data in
hexadecimal format. Non-programmers using Modbus or BACnet utilities (or those using
ViewPoint HD) may prefer to use decimal notation. This example works through the details of
converting information found in our user documentation (decimal) into a hexadecimal format
which should cover the highest level of complexity.
Entering Floating Point Data Using a Script
Consider the process of setting the CH1 CT Full Scale Rating to a value of 100.00 amps for a meter having
an element at slave address #1.
Step 1. Convert 100.00 into an IEEE 754 floating point format by entering the number in a conversion
utility.
a. Internet Utility. Example: https://www.h-schmidt.net/FloatConverter/IEEE754.html
b. ViewPoint HD
a) Internet Utility
The 16-bit floating point representation of 100.00 is shown below as 0x42C8 0x0000. These are
the required MSW and LSW register contents respectively.