Chinese
Spanish
Russian
Portuguese
French
German
Indonesian
Thai
Vietnamese
Arabic
Italian
Turkish
Polish
Japanese
Level Signal to PLC and DCS: Scaling and Configuration
The level transmitter measures correctly in the field, and the HMI still shows the wrong number. In almost every case the fault is not the instrument — it is a scaling mismatch between what the transmitter is sending and what the control system is configured to expect. Scaling is the translation layer between the 4–20 mA loop and the engineering units on the screen, and a single wrong assumption there shifts every reading downstream.
The two ends of the same scale
A 4–20 mA loop carries only one number: the current. What that current means is set at two places that must agree — the transmitter and the controller.
The transmitter is configured with a range: for example, 0 to 5 metres, where 4 mA = 0 m and 20 mA = 5 m. The controller is configured with a matching input range: the same 0 to 5 m, in the same units, against the same 4 and 20 mA points.
For a linear mapping, EU = LRV + (I − 4 mA)/16 mA × (URV − LRV). With a transmitter assigned to 0–5 m, 12 mA represents 2.5 m. A controller incorrectly scaled 0–10 m displays 5 m. A span mismatch gives a ratio error, while a zero mismatch also adds an offset.
The four classic scaling mistakes
Wrong zero or span at one end. Different endpoints produce gain error, offset or both. Match the assigned variable, units and 4/20 mA endpoints rather than assuming every mismatch is a constant ratio.
Wrong units. The transmitter is in metres and the channel is in feet, or in a percentage that assumes a different vessel height. The error is a constant conversion factor.
Wrong signal interpretation. A 4–20 mA channel treated as 0–20 mA shifts every reading by the missing live-zero offset, so zero reads as 20 %.
Square-root left on where it does not belong. Differential-pressure flow needs a square-root extraction; a level signal does not. A square-root function applied to a level loop distorts the reading most severely near the bottom of the range.
| Mismatch | Symptom | Check |
|---|---|---|
| Zero or span differ | Offset, ratio error or both | Reconcile variable, endpoints and datum |
| Units differ | Wrong numerical conversion | Record engineering units at both ends |
| 4–20 mA treated as 0–20 mA | 4 mA appears as 20% of the input scale | Select the correct channel range |
| Unwanted square root | Nonlinear distortion | Confirm whether the signal is level, flow or another assigned variable |
| Linearisation performed twice | Wrong volume despite correct level | Apply the validated conversion in one defined place |
Linearization for non-linear vessels
A level signal is linear with level, but the volume in many vessels is not. A horizontal cylindrical tank or a spherical tank has a non-linear relationship between level and volume: the same change in level corresponds to a different change in volume depending on where it is in the tank. To read volume rather than level, the control system (or the transmitter) applies a strapping table or a linearization curve that maps level to volume point by point.
If the transmitter assigns its analog output to volume and applies the table there, the current loop carries volume and the DCS must scale volume. If the transmitter outputs level and the DCS applies the table, the loop carries level. Some devices retain several digital variables, so verify the actual analog assignment. Apply the conversion once and document the source of the tank table.
Verifying the whole chain
The reliable way to confirm a scaled loop is to check it end to end:
Read the transmitter's configured range from its own menu or via HART.
Read the channel's configured range from the DCS engineering view.
Confirm the two agree in value and units.
Under the approved test and bypass procedure, inject or command known currents such as 4, 12 and 20 mA and verify the corresponding display, alarm and historian values. A current-injection test checks the receiving chain, not the sensor’s physical level accuracy. Verify configured fault currents and bad-quality handling separately where required.
A two-point check at 4 mA and 20 mA catches a zero/span error; a mid-scale check at 12 mA catches a square-root or linearization error that the endpoints would hide.
Questions engineers ask
Q1. Where should I do the range — in the transmitter or the DCS?
First identify the analog output variable: distance, level, volume or another quantity. Match the DCS scale to that assignment and its endpoints. Keep physical sensor setup distinct from any subsequent tank-volume conversion, and avoid applying that conversion twice.
Q2. Why does my level read 20 % when the tank is empty?
A 4–20 mA signal interpreted on a 0–20 mA scale is one possible cause. Check the measured current, assigned variable, physical zero, channel range and any forced or held value before changing the configuration.
Q3. Do I need square-root for a level signal?
No. Square-root extraction is for differential-pressure flow. A level loop is linear; leave square-root off.
Q4. How do I linearize a horizontal tank?
Use a strapping table (level vs volume points) generated from the tank's actual geometry, applied either in the transmitter or the DCS. A generic curve will be wrong for your vessel.
