Put a panel power meter on the compressor MCC feed, Modbus RTU, 600:5 CTs, reading it into a CompactLogix through a Modbus gateway. The meter's own display says 187 kW and 231 A, which is what the clamp on the feed says as well.
The register in the PLC comes in at 1874 for kW and the amps register at 2310. kVA, kVAR, every one of them ten times what the display shows. I swapped the word order in the gateway and got garbage in the millions, so I put it back, and the CT ratio in the meter checks out, 600 primary and 5 secondary. Bit stuck now.
Meter and clamp agree with each other and the PLC is the odd one out. is there a decimal setting in the meter that the display honours and the register doesn't?
Every register out by exactly ten is scaling, not word order. Word order gives you the garbage you already saw. Which meter is it, and does the register map have a scale or multiplier column next to the kW register?
Map page attached. There's a column I ignored, says "x0.1" against kW and against the current registers. And a register two addresses above the kW one called "scale factor". Not sure if that's the same thing twice.
The x0.1 column you skipped is the ten. The meter sends kW as an integer with one decimal implied, 1874 is 187.4 kW, and the current the same, 2310 is 231.0 A. The scale factor register tells you the exponent, and on a lot of meters it changes with the CT ratio, so read it and apply it in the ladder instead of hardcoding 0.1. Then a CT change later does not silently break your numbers.
Also check you are not one register off. 40001 against offset 0 has bitten me more than once, and the register next to the one you want is often the scale.
There's a writeup on getting meter data into the PLC here: https://plctr.com/using-plc-for-energy-management-and-conservation/
Read the scale factor register, it holds -1. Applied it in the ladder as ten to that power times the raw value. 187.4 kW, 231.0 A, matches the display.
The column I skipped on the map was the whole thing. Meter sends the number with a decimal implied and I was taking it raw, so reading the scale register and applying it is all it needed.
The kWh register is a 32 bit and has no x0.1 in the column, and I'm getting a tenth of what the display shows on that one, so I guess it's the other way round somewhere. Next week's problem. Thanks marco, scott.
kWh is a 32 bit with its own exponent. Different register, same game.