Home › Calibration

Pressure Transmitter Calibration: A Five-Point Field Procedure

Calibration · 10 min read

A calibration check compares what the transmitter outputs against a known reference pressure across its range. The goal is to record how the instrument actually performs ("as found"), correct it if needed, and record the final result ("as left"). This note walks through a simple five-point procedure for a 4–20 mA pressure transmitter.

What you need

A pressure source: hand pump, deadweight tester or pressure controller.
A reference pressure standard. It should be clearly more accurate than the device under test. A test accuracy ratio of about 4:1 is a common rule of thumb.
A meter or loop calibrator to read the mA output.
A HART communicator if you need to trim a smart transmitter.
The instrument datasheet, its configured range (LRV/URV) and the allowed tolerance.

Figure 1 — A typical bench or in-situ setup. The pump pressurizes the transmitter and the reference standard together. The loop calibrator powers the loop and reads the output.

1. Prepare and isolate

Get permission from operations and put the loop in manual or bypass as your site requires. Isolate the transmitter from the process with its block valve or manifold, using your site's valve sequence. Then safely vent the trapped pressure before connecting your test equipment.

2. Choose the test points

Use five points across the span: 0, 25, 50, 75 and 100 %. For a transmitter ranged 0–10 bar:

% of span	Applied pressure	Expected output
0 %	0.0 bar	4.00 mA
25 %	2.5 bar	8.00 mA
50 %	5.0 bar	12.00 mA
75 %	7.5 bar	16.00 mA
100 %	10.0 bar	20.00 mA

3. Record the "as found" readings

Do not adjust anything yet. Apply each point going up (0 → 100 %), then come back down (100 → 0 %). Let the reading settle at each point before recording it. Approaching every point from both directions shows hysteresis: a difference between the upscale and downscale readings at the same pressure.

4. Calculate the error

Error is usually expressed as a percentage of span. For a 4–20 mA output the span is 16 mA:

Error (% of span) = (I_measured − I_expected) ÷ 16 mA × 100

Example: at 50 % the expected output is 12.00 mA and you measure 12.08 mA:

Error = (12.08 − 12.00) ÷ 16 × 100 = +0.50 % of span

Figure 2 — Plotting the as-found errors. Every point is inside ±0.5 %, so this transmitter passes. The gap between the upscale and downscale lines is the hysteresis.

Compare every point with the tolerance set for that loop, for example ±0.5 % of span. If every point is within tolerance, record the result as found and as left, and the calibration is complete.

Example calibration record

Point	Applied	Expected	Up (mA)	Down (mA)	Max error	Result
0 %	0.0 bar	4.00	4.008	4.024	+0.15 %	Pass
25 %	2.5 bar	8.00	8.019	8.053	+0.33 %	Pass
50 %	5.0 bar	12.00	12.032	12.067	+0.42 %	Pass
75 %	7.5 bar	16.00	16.040	16.061	+0.38 %	Pass
100 %	10.0 bar	20.00	20.048	20.048	+0.30 %	Pass

5. Adjust if needed

Figure 3 — How the two most common errors look. A zero error shifts the whole line; a span error changes its slope.

Zero error: every point is shifted by about the same amount. Fix the 0 % point with a zero trim.
Span error: the error grows toward the top of the range. Fix it with a span (upper) trim after the zero is correct.
Linearity or hysteresis error: a trim usually cannot fix this. It may point to a damaged sensor, an overpressure event or a mechanical problem.

On smart (HART) transmitters, there is a difference between sensor trim and re-ranging. A sensor trim corrects how the transmitter reads pressure. Re-ranging only changes which pressures map to 4 and 20 mA. Re-ranging does not correct a measurement error.

6. Record the "as left" readings

Repeat the full five-point, up-and-down check after adjusting, and record it as the "as left" result. Include the date, the reference standard's ID and certificate due date, the technician's name, and the tolerance applied.

7. Return to service

Remove the test equipment and refit any plugs or vents. Return the manifold to its in-service position using your site's sequence. Confirm the reading in the control system agrees with the process before handing the loop back to operations.

Tip: the as-found readings are the most valuable part of the record. Over several calibrations they show how quickly an instrument drifts. That history tells you whether the calibration interval can be extended or should be shortened.