Let us now examine the general properties of the controller shown in Figure 2.

To review: (1) Inputs to the controller are a measured indication of both the controlled variable and a setpoint representing the desired value of the

variable, expressed in the same fashion as the measurement; (2) the controller output is a signal representing action to be taken when the measured

value of the controlled variable deviates from the setpoint. The measured indication of a variable is denoted by b, and the actual variable is de-

noted by c. Thus, if a sensor measures temperature by conversion to resistance, the actual variable is temperature in degrees Celsius, but the

measured indication is resistance in ohms. Further conversion may be performed by transducers or transmitters to provide a current in mA, for

example. In such a case, the current becomes the measured indication of the variable.

Error

The deviation or error of the controlled variable from the setpoint is given by

e=r- b
where

e = error

b = measured indication of variable

r = setpoint of variable (reference)

Equation (2) expresses error in an absolute sense, or in units of the measured ana-
log of the control signal. Thus, if the setpoint in a 4- to 20-mA range corresponds to
9.9 mA and the measured value is 10.7 mA, we have an error of -0.8 mA. Obviously,
this current error has little direct meaning unless related to the controlled variable.
We could work back through the loop and prove that it corresponds to a flow rate of
1.1 m3/h, for example. This would show the significance of the error relative to the
actual process-control loop.
To describe controller operation in a general way, it is better to express the error as
percent of the measured variable range (i.e., the span). The measured value of a
variable can be expressed as percent of span over a range of measurement by the
equation
c - cmin
CP = * 100
cmax - cmin
 cp = measured value as percent of measurement range
c = actual measured value
cmax = maximum of measured value
cmin = minimum of measured value

The previous equation is in terms of the actual measured variable, c, but the same
equation can be expressed in terms of the measured indication, b. It is necessary only
to translate the measured minimum and maximum to bmin and bmax.
To express error as percent of span, it is necessary only to write both the setpoint
and measurement in terms of percent of span and take the difference according to
Equation (2). The result is

r- b
eP = * 100

bmax - bmin
(1)
ep = error expressed as percent of span

You can see the convenience of using a standard measured indication range like 4 to
20 mA, because the span is always 16 mA. Suppose we have a setpoint of 10.5 mA and
a mea- surement of 13.7 mA. Then, without even knowing what is being measured, we
know the error is

10.5 mA - 13.7 mA
ep= 20 mA - 4 mA * 100
ep = -20%

A positive error indicates a measurement below the setpoint, and a negative error
indicates a measurement above the setpoint.

Control Parameter Range

Another range is associated with the controller output. Here we assume the final control el- ement
has some minimum and maximum effect on the process. The controller output range is the
translation of output to the range of possible values of the final control element. This range is also
expressed as the 4- to 20-mA standard signal, again with the minimum and maximum effects in
terms of the minimum and maximum current.
Similarly, in computer-based control, the output will range over all states of the n-bit output.
Generally, all 0’s are the minimum output and all 1’s the maximum. These numbers do not
necessarily represent the minimum and maximum of the final control element, how- ever. We may
wish a valve never to be fully closed, for example; therefore, all 0’s might represent some
percentage of full open.
Often, the output is expressed as a percentage where 0% is the minimum controller output and 100%
the maximum (obviously). Thus, in the example of Figure 1, the valve in the fully open position
corresponds to a 100% controller signal output. Often, however, the minimum does not correspond
to zero effect. For example, it may be that the steam flow should never be less than that flow which
results with the valve half open. In this case, a 0% minimum controller corresponds to the flow rate
with a half-open valve.
The controller output as a percent of full scale when the output varies between spec- ified limits is
given by
u - umin
p= * 100
umax - umin (1)
where p = controller output as percent of full scale
u = value of the output
umax = maximum value of controlling parameter
umin = minimum value of controlling parameter
 ELECTRONIC CONTROLLERS

In the following treatment of electronic methods of realizing controller modes, emphasis is on the use of op amps as the primary circuit element.

Discrete electronic components also are used to implement this function, but the basic principles are best illustrated using op amp circuits. Op amp

circuits other than the ones described also can be developed.

Error Detector

The detection of an error signal is accomplished in electronic controllers by taking the dif- ference between voltages. One voltage is generated by

the process signal current passed through a resistor. The second voltage represents the setpoint. This is usually generated by

a voltage divider using a constant voltage as a source.

An example is shown in Figure 2. We assume a two-wire

system is in use so that the current drawn from the

floating power supply is the 4- to 20-mA signal current.

The signal current is used to produce a voltage, IR,

across the resistor, R. This is placed in series

opposition to a voltage, Vsp, tapped from a variable resistor, Rsp, connected to a constant positive source, V0. The result is an error voltage,

Ve = Vsp - IR. This is then used in the process controller to calculate controller output.

An error detector also can be made from a differential amplifier. Such a system can be used only if the current from the transducer is referenced

to ground. Figure 3 shows one typical configuration. The sensor signal current passes to ground through RL, provid- ing a signal voltage, Vm = IRL. The

differential amplifier then subtracts this from the set- point voltage.

Proportional Mode

Implementation of this mode requires a circuit that has a re- sponse given by

KP = proportional gain
p = KPep + p0
ep = error in percent of variable range

where p = controller output 0–100% p0 = controller output with no error

If we consider both the controller output and error to be expressed in terms of voltage, we see that above Equation is simply a summing amplifier.

The op amp circuit in Figure shows such an electronic proportional controller. In this case, the analog electronic equation for the output voltage is

V out P e + V0
=G V

Where

Vout = output voltage GP = R2 /R1 = gain Ve = error voltage

V0 = output with zero error
The design of a proportional controller calls for specification of the proportional gain described by KP in Equation that expresses the percent of

output for an error of 1% of the measurement range. Alternatively it could be described as the proportional band, PB 100/KP. This must now be

expressed in terms of the voltage gain, GP , in Equation (3). The relationship between GP and K is given by,

where

Vout the range of output voltage

Vm the range of measurement voltage

A temperature controller controls temperature from 100° to 200°C. A sensor provides an output of 2 to 8 V for

this temperature range. The controller output drives a heater with an output of 0 to 5 volts. What circuit

gain is needed if the controller of Figure 7 is to be used with a proportional gain of 4%/%?

The range of measurement voltage will be Vm 8V - 2V = 6 V. The output range is 5 volts. Thus, the

circuit gain will be,