Instrumentation basics · 1.2

The instrument loop: sensor, transmitter, controller, final element

25 October 2026 · 9 min read · ISA 5.1, ISA 5.4, IEC 60381-1

In short

A control loop has four parts in a ring. The sensor feels the process, the transmitter sends a signal, the controller decides, and the final element acts.
The signal between them is usually 4–20 mA, where 4 mA is 0 % and 20 mA is 100 % of the range.
Formula: I = 4 + 16 × (PV − LRV) / (URV − LRV).
The commonest loop fault is not a broken instrument. It is a range that differs between the transmitter and the control system.

Every control loop, whether it holds a level, a pressure or a temperature, has the same four parts in the same order. Once you can trace one loop from the pipe to the valve and back, you can read any loop on any project.

Where this matters

Each part of the loop is usually specified on a different document, bought from a different vendor and configured by a different person. The transmitter is set up by the instrument technician, the controller by the control system engineer, and the valve by the valve vendor. A loop works only when all four agree on the range, the signal, the action and the fail-safe position. Loop checks at commissioning exist to prove that agreement.

How it works

1. Sensor (primary element)

The sensor is in contact with the process, or close to it, and changes in a measurable way. A diaphragm deflects with pressure. An RTD's resistance rises with temperature. An orifice plate creates a pressure difference with flow. On P&IDs a sensor carries the letter E, as in TE or FE.

2. Transmitter

The transmitter converts the sensor's raw output into a standard signal, scaled to a chosen range. It has a lower range value (LRV) and an upper range value (URV). Analogue current is still the most common signal:

I = 4 + 16 × (PV − LRV) / (URV − LRV) [mA]

Here PV is the process value, and LRV and URV are in the same units as PV. Because zero is 4 mA and not 0 mA, a broken wire (0 mA) can be told apart from a true zero reading. The 4 mA also powers the transmitter in a 2-wire loop.

3. Controller

The controller, usually a function block in the DCS, converts the current back into engineering units using its own range setting. It compares the result with the set point and calculates an output (OP) from 0 to 100 %. It also has an action. Direct action means that a rising PV makes the output rise. Reverse action means that a rising PV makes the output fall. The action must match the valve and the process.

4. Final element

The output goes out as another 4–20 mA signal to a valve positioner. The positioner moves the valve actuator to the matching travel. The valve changes the flow, the flow changes the process, and the sensor sees the change. That closes the loop.

Every final element also has a fail position: fail open, fail closed or fail last. This is what the valve does when it loses air or signal. It is chosen for process safety, not for convenience.

What the standards and practice say

Item	Reference	Notes
4–20 mA analogue signal	IEC 60381-1, ISA 50.00.01	Current is used because it is not affected by cable resistance within the loop's limits
Failure signal levels	NAMUR NE43	≤ 3.6 mA or ≥ 21 mA mean a failed transmitter, not a real value (article 7.2)
Loop symbols and tag letters	ISA 5.1	Shows which part is the sensor, transmitter, controller and valve (article 1.3)
Loop diagrams	ISA 5.4	Shows every terminal, cable and junction box in the loop (article 8.4)
Digital signals on top of 4–20 mA	HART	Adds diagnostics and configuration without changing the analogue value (article 7.3)

Worked example: tracing a pressure loop

Here is a pressure control loop on a vessel, PIC-201. Pressure is controlled by a valve on the gas outlet.

Item	Value
Transmitter PT-201 range	0–10 barg
Process pressure	6.5 barg
Set point	6.0 barg
Controller output	55 %
Valve PV-201	Fail closed, air to open

Step 1, transmitter signal. I = 4 + 16 × (6.5 − 0) / (10 − 0) = 4 + 10.4 = 14.4 mA.

Step 2, controller reading. With the DCS range also set to 0–10 barg, it reads (14.4 − 4) / 16 × 10 = 6.5 barg. Correct.

Step 3, action. The pressure is above the set point, so the valve must open more to vent gas. A rising PV must give a rising output, so the controller is direct-acting.

Step 4, valve signal. An output of 55 % gives 4 + 16 × 0.55 = 12.8 mA to the positioner, and the valve moves to 55 % travel.

Now the common fault: a range mismatch

Say the transmitter was delivered ranged 0–16 barg, but the DCS was configured for 0–10 barg.

Case	Transmitter sends	DCS shows	Result
Ranges match (0–10 / 0–10)	14.4 mA	6.50 barg	Correct
Transmitter 0–16, DCS 0–10	10.5 mA	4.06 barg	Reads 2.4 bar low

The controller now believes the pressure is 4.06 barg, well below the set point. It closes the valve, and the real pressure keeps rising until a relief valve or a trip acts. Nothing is broken. Each device is working exactly as configured.

What this example shows

The range lives in two places, the transmitter and the control system, and both must match the datasheet.
Controller action depends on the valve's fail position and its place in the process. Getting it backwards gives a loop that runs away instead of correcting.
A loop check injects known values, such as 0, 25, 50, 75 and 100 %, and confirms the reading at the DCS. It catches exactly this kind of fault before start-up.

When it fails: the fixes in order

Check the range in all three places. These are the datasheet, the transmitter configuration and the DCS tag. Correct the one that differs.
Check the signal. Measure the loop current. Below 3.6 mA or above 21 mA, the transmitter is reporting a failure, so look at the instrument, not the process.
Check the action and the fail position. Stroke the valve from the DCS and confirm the direction.
Check the wiring and power. A 2-wire loop needs enough supply voltage to cover the transmitter, the cable and the input resistance (article 7.1).
Only then tune the controller. Tuning cannot fix a wrong range or a wrong action.

Common mistakes

Changing the transmitter range in the field without updating the DCS and the datasheet.
Setting the controller action without checking whether the valve is air-to-open or air-to-close.
Choosing a fail position for convenience instead of from the process safety review.
Treating 0 mA as zero process value instead of a broken loop.
Using a 0–100 % range on the DCS and leaving the engineering units for later.
Skipping the loop check because "the instruments were calibrated in the workshop".

Checklist

Sensor, transmitter, controller and final element are all identified with the same loop number.
LRV, URV and units match on the datasheet, the transmitter and the DCS.
The signal type is stated: 4–20 mA, HART or fieldbus.
The failure signal direction (upscale or downscale) is set and handled in the DCS.
The controller action is set and checked against the valve's action.
The valve fail position comes from the safety review and is on the datasheet.
Loop power is enough for the transmitter, the cable and the input card.
A loop check at 0, 25, 50, 75 and 100 % is recorded at commissioning.

Quick reference

Signal: I = 4 + 16 × (PV − LRV) / (URV − LRV) mA
Back to units: PV = LRV + (I − 4) / 16 × (URV − LRV)
Key points: 4 mA = 0 %, 8 mA = 25 %, 12 mA = 50 %, 16 mA = 75 %, 20 mA = 100 %
Failure (NE43): ≤ 3.6 mA or ≥ 21 mA
Direct action: PV up → output up. Reverse action: PV up → output down.