AB PIDE block

Hello all. I'm wondering when programming a PID loop, do you typically bring the SP and PV in as a scaled value or as a raw value? it would seem that there would be better resolution in a raw value but harder to make sense of it when troubleshooting,etc. I have seen it done both ways-just wondering opinions. Or is my thinking off and since it's a REAL it doesn't matter?

I used to use scaled engineering units for PV (often as an int) and CV was usually an INT with 0-10000 for 0-100%. These days I scale to per-unit so that 1 is the nominal PV. using per unit for the PV gets the maximum resolution from lag filter with long time constant.

My use case is hydro power plant where people like to see the head and tail water measured in metres above sea level for easy calculation of the hydraulic head across the turbine and turbine efficiency. So while I display the water levels in m ASL on the HMI I take 0 as the head pond reference elevation. The SP might be -0.5m. Using -8 to + 2 as the CV with normal = 0 then lets me use a 60s lag filter and get higher resolution water level measurement than if I was using the normal value of 1000 m ASL in to the same lag filter it would not display as many decimals.

Using per unit values also makes it so the gains for speed and level control loops at various plants are on the same scale so I can see how the physical parameters of each plant affect their controlability.