Compression algorithm, Appendix, Appendix d – Rockwell Automation 1426 PowerMonitor 5000 Unit User Manual
Page 379

Rockwell Automation Publication 1426-UM001F-EN-P - November 2013
379
Appendix
D
PowerMonitor 5000 Waveform Capture and
Compression (M6 model)
Waveform recordings in the PowerMonitor 5000 M6 model consist of a series of
cycle-by-cycle magnitude and angle data for each spectral component (harmonic)
from DC through the 127th harmonic. To reduce the size of waveform records
without losing significant resolution, the data is compressed before writing to the
waveform file. To display the record as a waveform, the file data must be
decompressed, and then an inverse FFT performed to obtain a series of time-
domain voltage and current data that can then be plotted in a graphic format.
Compression Algorithm
Three types of floating point number representations are used, with 32, 16 and
12 bits. The formats are summarized in the table.
The table below defines how compression is applied to magnitude and angle
values of specific harmonic orders.
The various number encodings are packed into the file in the following way:
Type
Total bits
Bits
precision
Sign bits
Exponent
bits
Significand
bits
Exponent
bias
IEEE 754
Single
32
24
1
8
23
127
16 bit
encoded
16
12
1
4
11
TBD
12 bit
encoded
12
8
1
4
7
TBD
Data / encoding
32-bit
16-bit
12-bit
Magnitude
DC thru 15th
-
16th thru 127th
Angle
-
DC thru 15th
16th thru 127th
Table 196 - 32-bit (IEEE 754)
Byte offset 0
Byte offset 1
Byte offset 2
Byte offset 3
Low byte
Next lowest byte
Next highest byte
High byte
Table 197 - 16-bit Encoding
Byte offset 0
Byte offset 1
Low byte
High byte