Process Variable Measurement

Aug. 1, 2011
Newer Controllers and Smart I/O Now Make Analog Variables Such as Pressure, Temperature, Flow and Level Easier to Measure
By Dan Hebert, PE, Senior Technical Editor

OEM machine and skid builders often are required to measure analog variables such as pressure, temperature, flow and level—and use these variables for closed-loop control. Newer controllers and smart I/O now make these tasks simpler, cheaper and quicker.

For example, an upgrade to a new controller often requires interface to existing analog sensors via existing wiring. Because of ground loop and noise issues, the new controller sometimes can't interface directly to the existing sensors, so an interface component is required.

For one such project, Phoenix Contact ( says it provided signal isolators between the controller and the sensors to isolate and break the ground loop. These isolators can handle more than 200 different types of inputs via DIP switch settings, so the installation required only one universal isolator type.

In another case, a cleanroom system manufacturer found that existing analog transmitters connected to field sensors weren't able to limit their analog output to 20 mA. This created upstream PLC signaling errors that required in-room troubleshooting. To solve the problem, Wago ( reports that it provided an isolation amplifier with signal clipping functionality to limit output when the input from the sensor exceeded 20 mA.

Another type of analog signal conditioner, also from Wago, provides input circuit protection via a resettable fuse. After an overcurrent condition is cleared, the fuse automatically resets itself, saving the task of field replacement.

Sukup Manufacturing ( in Sheffield, Iowa, says it used a PLC and I/O modules from Phoenix Contact to provide closed-loop temperature control for its new grain dryer. Grain dries faster at higher temperatures, but can be scorched and damaged by excessive heat. RTDs monitor plenum temperature and feed this signal to analog PLC inputs. Sukup uses tuned PID loops to control the temperature to ±3 °F of setpoint by sending analog output signals from the PLC to electronic modulating valves that control gas flow to the heaters.
In another closed-loop control application, PID loops weren't sufficient to provide the level of desired control, so more sophisticated and advanced control techniques were required. Industrial Furnace (IFCO,, headquartered in Rochester, N.Y., makes multiple-hearth furnaces. Runaway conditions are a concern with furnace applications because of their high thermo mass. To get better control of the process loops, Industrial Furnace tells us it implemented the PlantPAx process automation system from Rockwell Automation (

IFCO used an internal model control (IMC) function block to replace PID control for some of the control loops. For others, a coordinated control (CC) function block replaced PID and provided multi-control variable time-based control.

In all, there are 12 control loops. These loops are used in various configurations to control temperature via combustion air valves, water flow control valves, and variable frequency drives that control fan speed and thus temperature.

In each of these loops, simple PID control wasn't able to provide the desired minimal variance between the set point and the process variable. This was mainly because of the furnace's high thermo mass, a condition that made for long dead times. The dead time is the time it takes the PV to change after the control loop's output changes. The longer the dead time, the more difficult it is to control the loop with PID.

The two types of advanced control used depended on the particular characteristics of each of the 12 control loops. IMC, which is an algorithm that models the process as a first-order lag with dead time, addresses loops with dead time particularly well. CC was used on other loops. With CC, up to three outputs are modeled as a first-order lag with dead time. Each model relates how a single PV relates to the multiple controlled variable outputs, and the additional outputs can have targets that the algorithm drives to for optimum control.