Chinese
Spanish
Russian
Portuguese
French
German
Indonesian
Thai
Vietnamese
Arabic
Italian
Turkish
Polish
Japanese
Radar Level Transmitter Commissioning: A Field Checklist That Catches Real Errors
Commissioning verdict. Radar level transmitter commissioning is not complete when the display shows a plausible number. It is complete when the mechanical reference, electrical limits, echo evidence and independent level checks agree. A transmitter can be configured to 100% output on an empty tank and still pass a careless loop test, so the sequence below starts with the installation rather than the keypad.
The record to open before touching the configuration
Create one commissioning record per tag. If the installer, PLC programmer and instrument engineer each keep a different version, the next maintenance shift inherits three versions of the truth.
| Record field | What to enter | Why it matters later |
|---|---|---|
| Tag, serial number and firmware | Copy from the nameplate and device menu | Parameter meanings can change between models |
| Measurement reference | Flange face, thread end or antenna reference | Every offset and empty distance depends on it |
| Tank geometry | Drawing number, height, nozzle and stilling-well details | Explains echoes that appear months later |
| Empty and full distances | Values in the device’s distance convention | Prevents the common level/distance reversal |
| Materials and area approval | Antenna, seal, certificate and temperature class | Confirms the delivered unit matches the service |
| As-found and as-left files | Export both if the tool permits it | Separates inherited errors from commissioning changes |
| Echo curves | Empty, known level and agitated condition where relevant | Baseline evidence, not just a screenshot for the file |
Photograph the nozzle before insulation or cladding is closed. A later “false echo” investigation is much easier when the team can see what is actually inside the tank connection.
Mechanical checks that software cannot repair
Confirm verticality and the antenna reference. A tilted antenna changes the return path and can move the strongest reflection toward a nozzle wall. Use the manufacturer’s permitted alignment range rather than assuming the flange is square to the liquid surface.
Inspect the nozzle with a light or camera only under the approved isolation, access and permit procedure. Look for a protruding gasket, weld lip, rust scale, a blind flange left in place, or an antenna recessed behind a long restriction. Do not enter a confined space to perform this inspection.
Check clearance from the fill stream and agitator. A stable empty-tank echo does not prove the unit will work during filling.
Verify the process seal and cable entries. The transmitter’s ingress rating does not protect a loop with an open gland or a conduit that drains water into the housing.
For a guided-wave probe, confirm suspension and bottom clearance. Record probe length, weight or anchor arrangement, and any restriction on product contact.
If a nozzle problem is found, stop. Suppressing its echo in software may hide the symptom while also suppressing a real level near the top.
Prove the loop with the actual supply and barriers
A loop that works on the bench can fail at 20 mA in the field because loop voltage falls as current rises. Use the device’s minimum terminal voltage, not a generic value. An illustrative non-hazardous calculation is:
Available load resistance = (24.0 V − 12.0 V) ÷ 0.022 A = 545 Ω
Under the assumed values, 545 ohms is the total resistive allowance at 22 mA. Subtract an existing 250-ohm communication load only if it is actually required and not already included, together with other resistive loads. If those other loads are 80 ohms, about 215 ohms remains. For barriers or isolators with a specified voltage drop, include that drop in the voltage equation rather than treating every device as a resistor. Recalculate using the actual transmitter, supply and associated-equipment limits.
Then verify:
polarity, shield termination and the drawing’s grounding arrangement;
the 4 mA and 20 mA points at the PLC input, not only on the transmitter display;
HART communication with the final loop resistance installed;
the failure current and the PLC’s response to a device fault;
engineering units and decimal places through the whole chain.
Do not use a loop calibrator to “correct” a level error. A current trim fixes the output chain; it does not fix a wrong empty distance.
Configure in an order that preserves traceability
Change one group of parameters at a time and record the result. A practical order is units and reference, tank distances, application or medium type, output mapping, damping, then echo evaluation. This prevents a technician from tuning the echo algorithm to compensate for a basic dimensional mistake.
| Parameter | Field question before saving | Typical mistake |
|---|---|---|
| Empty distance | Is this from the stated reference to the tank zero? | Entering tank height instead of measured distance |
| Full distance | Does 100% mean the operating full point or the physical overflow? | Confusing working capacity with total capacity |
| Blocking distance | Does it protect the near zone without hiding a required alarm level? | Copying the largest available value |
| Damping | How fast must the control or alarm respond? | Using damping to hide an installation fault |
| False-echo suppression | Was it learned under a known, documented condition? | Mapping with product at an unknown level |
| Output and fault current | Do the PLC and safety philosophy agree? | Leaving a default NAMUR value untested |
A false-echo map is evidence about one tank condition. Save the curve before mapping, the map itself and a curve afterward. If the tank has removable internals, agitators or a heating coil, repeat the review under the operating states that matter.
The wet test: three points plus one dynamic condition
Compare against an independent reference whose own uncertainty is understood: a calibrated dip tape where permitted, a sight gauge with verified taps, or another validated reference. A weigh scale or metered transfer can support a level comparison only through the required geometry and, where relevant, density conversion with its uncertainty included. A second reading derived from the same transmitter is not independent.
At low, mid and high level, record the reference, transmitter distance, transmitter level, PLC value and difference. Wait for the process to stabilize, but do not increase damping just to make the comparison easier. Also observe filling, emptying or agitator start-up. Dynamic service often exposes spray, vortex and multiple echoes that a static test misses.
A simple sign check prevents a bad offset correction. Suppose the independent reference says 3.20 m and the transmitter says 3.10 m. The transmitter is 0.10 m low. Before adding an offset, confirm whether the error is constant at the other points. If the differences are 0.10, 0.32 and 0.55 m, it is not a single offset; investigate the reference, density-independent measurement path, nozzle, scaling or echo selection.
Acceptance criteria to agree before the test
The project specification should define acceptance before the vendor and contractor are standing at the tank. Include maximum permissible error at each point, allowed repeatability during a stable condition, PLC scaling error, alarm operation, fault-current operation, required echo-curve files and who signs the record. Do not invent an acceptance band on site by widening it until the readings pass.
Questions from the commissioning shift
Q1. Can I commission before product is available?
Yes, for mechanical, electrical and parameter checks plus an empty-tank baseline. Final level accuracy and dynamic behavior still need verification with product or another defensible reference.
Q2. Should I always perform a false-echo mapping?
No. Follow the device method and the condition it assumes. Mapping over an uncertain or changing surface can teach the transmitter to reject a genuine echo.
Q3. Why does the local display agree with the dip tape but the PLC disagree?
Check the 4–20 mA range, PLC raw counts, square-root or linearization settings, integer truncation and whether another system applies a second offset.
Q4. Is a 4 and 20 mA simulation enough for handover?
It proves part of the output path only. It does not verify the measurement reference, tank geometry, echo selection or dynamic performance.
Q5. What should be handed to maintenance?
The signed record, final parameters, as-left echo curves, loop and nozzle drawings, certificates, independent-reference results and a clear list of unresolved deviations.
