Improper Sample Intervals - Johnson Controls - Metasys - LIT-12011147 - Software Application - Controller Configuration Tool - 15.0

Controller Tool Help

Johnson Controls
Product name
Controller Configuration Tool
Document type
User Guide
Document number
Revision date

PRAC+ is able to tune a PID controller while the loop is in control with relatively little information. However, a general idea of the dynamics of the system is necessary to choose a proper sampling interval. PRAC+ works well with sampling intervals between 1/40 and one times the time constant of the system, with the optimal sampling interval being around an eighth of the time constant. If the process is sampled too slowly, PRAC+ must either tune the loop to a sluggish control state to compensate for the long time between samples, or interpret the oscillations as noise, and not tune at all. If the process is sampled too fast, after 60 or 75 samples, PRAC+ may misinterpret a response as sluggish because it did not see the information contained in the full pattern. Figure 1 shows an example of poor control due to slow sampling, and Figure 2 shows an example of poor control due to fast sampling.

Note: The process variable analog inputs have a process ID defined as well. Make sure to select the correct process ID for the analog input so the sample times are representative of the system and work with the interval of the PID. Analog inputs have a sampling period and a filter. If this input is not chosen correctly, sampling may occur too slowly or too quickly. For example, if you are performing DA-T control and the analog input is set to ZN-T, the PID samples every 12 seconds but the process variable only updates every 30 seconds, which leads to bad control. On the other hand, if the analog input is set to duct static pressure control and the control is a zone loop, the analog input filter cutoff frequency is too high and high frequency disturbances may be passed to the PID, leading to poor tuning and control.
Figure 1. Poor PID Tuning Due to Slow Sampling

Figure 2. Aggressive PID Tuning Due to Fast Sampling

Process IDs supply optimal PID values including interval for normal systems from within a class of control loops. However, a single process ID cannot cover all possible control loops in a class. Zones, for example, may have different time constants based on the size of the zone. A duct static pressure loop may have different effective time constants due to the ramp time of the variable frequency drive (VFD) controlling the fan. If the PID is performing poorly, you may need to change its interval back to within its tunable range to get an estimate of the time constant. See Process IDs and Performing a Step Test for more information.