Chinese
Spanish
Russian
Portuguese
French
German
Indonesian
Thai
Vietnamese
Arabic
Italian
Turkish
Polish
Japanese
Load cell calibration establishes how an installed weighing system responds to known loads. Before adjusting zero or span, keep a record of what the scale actually does: whether it returns to zero, whether the same load gives the same result, and whether moving that load changes the reading. A single successful span adjustment can hide a binding support or an unevenly loaded platform. This procedure separates those faults from errors that an adjustment can correct.
Start with the quantity the scale must measure
Define the working range and the smallest load change that affects a production decision. A vessel may hold several tonnes but dispense much smaller batches. Demonstrating a reasonable full-load reading does not establish that the batch increment is measured adequately. State whether the acceptance limit applies to gross weight, net contents or the difference between two readings.
Identify the final value used by the operator or control system. The indicator, PLC and production record should agree on units and tare treatment. If a transmitter sends gross weight while the PLC subtracts a stored tare, both calculations belong in the test. Do not assess only the local display when a different value controls filling.
Record the error pattern before changing settings
Save the existing calibration coefficients, zero offset, filters and junction-box settings. Note the empty mechanical configuration, including hoses, platforms and fixtures. An empty tank with a connected transfer hose is not necessarily equivalent to the same tank with that hose disconnected. Record temperature and the state of equipment that can transmit force into the scale.
| Observed result | Check first | Why span alone fails |
|---|---|---|
| The same load changes with position | Load distribution, mounting freedom and corner balance | One multiplier cannot correct different errors at different positions. |
| Zero changes after a valve operates | Pipe movement, pressure thrust and hose routing | The added force depends on process state. |
| Reading drifts without a load change | Temperature, creep, moisture and excitation stability | The test has no stable baseline. |
| Error grows consistently with load | Span, reference load and unit conversion | Adjustment may help, but confirm repeatability and unloading behavior first. |
These patterns identify checks, not diagnoses by themselves. A moisture-affected connection can produce an apparently mechanical symptom, and a cable pulled tight against the structure can disturb an electrically healthy sensor. Change one condition at a time and retain the observations.
Prepare reference loads and a repeatable loading method
Choose reference equipment with documented traceability and uncertainty suitable for the required acceptance limit. A certificate without the applicable load range or identification is not enough. Check how the reference will be placed, supported and removed. A load touching a wall or resting partly on a lifting device does not apply its full weight to the intended measuring structure.
For a large tank, full-range test weights may be impractical. A planned substitution or material-transfer method can extend the test, but transfer losses, retained material and the reference measurement then contribute to uncertainty. Define those contributions before the test. Do not label an unmeasured delivery quantity a calibration reference.
Use loading points that cover the actual working range, including the range important to batching. Specify the settling criterion and record the time at each reading. A stable-looking display after heavy filtering may still lag the applied load. Keep the filter setting unchanged during a comparison unless its effect is the subject of the test.
A load cell calibration sequence that preserves useful evidence
- Inspect the complete load path. Remove unintended contact and check restraints, mounting faces, pipe connections and cable strain. A competent technician should carry out electrical tests with the equipment isolated as required by its instructions.
- Allow the electronics to stabilize according to their instructions. Record the unloaded indication before applying a zero adjustment. On a vessel, establish that the stated empty condition is actually present.
- Apply the planned increasing loads and record the reference, indication and load position. Do not adjust after every point: that would erase the original error pattern.
- Remove load in planned steps. Compare loading and unloading readings at the same reference values, then record the return to zero.
- Repeat a relevant load point using the same placement procedure. For a platform or weighbridge, perform the position tests required by the application, not only a centre-load test.
- Make permitted adjustments only after assessing the as-found results. Repeat the affected tests and record the as-left results separately.
For each reading, indication error equals indicated value minus reference value. If a 500 kg reference produces 500.4 kg, the calculated error is +0.4 kg. This example is not a tolerance or a measured METRAVON result. The site's acceptance limit and the test uncertainty determine whether that result is acceptable.
Use the sensor data to check the test setup
The METRAVON MV-LC0190 has a nominal sensitivity of 0.5 mV/V at its 1 kg rating, compared with 1.0 mV/V for the 5, 20 and 50 kg versions. At 10 V excitation, the calculated full-scale signal changes are 5 mV and 10 mV respectively. Using one amplifier setup for both without checking the calibration can create a plausible but incorrect result.
The MV-LC0030 is a high-capacity compression column design. Its nominal 2 mV/V output gives a calculated 20 mV change at rated load with 10 V excitation. That relationship is useful for checking signal scale, but it is not a substitute for applying a reference load through the assembled scale.
For multi-support vessels, read the tank weighing installation guide before trimming unequal signals. Support reactions may genuinely differ. Equalizing millivolt readings without accounting for capacity, sensitivity and actual loading can conceal a mechanical problem.
Close the test with a usable result
Keep the raw readings, reference identification, test positions, environmental observations and settings together. State the tested range and any restrictions. If only part of the working range was tested, do not imply that the remaining range has been demonstrated. Any extrapolation needs a stated basis and an assessment of its effect on the decision.
For a service quotation, identify the scale capacity, working loads, number and model of sensors, access for test loads, indicator and final PLC signal. Include the observed fault pattern and available reference equipment in your calibration enquiry. Access, outage time, load handling and the test method can affect the work more than the sensor's purchase price.
Calibration questions from the workshop
Q1. Is pressing zero the same as calibration?
No. Zeroing assigns an offset at one condition; it does not establish the response to applied load. If buildup, a tight pipe or a damaged support caused the offset, zeroing can hide the symptom while the error returns at another temperature or load.
Q2. Can a factory certificate replace the installed test?
A sensor certificate describes the identified sensor under its test conditions. The installed result also depends on mounting, load distribution, wiring and the indicator. Use the certificate to support traceability and configuration, then test the assembled system for the required operating decision.
Q3. Why test a platform away from its centre?
A central load may distribute force evenly and conceal a weak corner, binding restraint or mismatched channel. Position tests examine conditions closer to normal use. Use defined locations and the same reference so changes in placement can be distinguished from changes in load.
Q4. Should the PLC be included in calibration checks?
Include it when its value controls or records the process. Compare the indicator and PLC at more than one load, including tare and sign behavior. The load-cell-to-PLC guide explains where a second scaling operation can introduce an error.
Q5. How often should the scale be recalibrated?
Use service history, load severity, required uncertainty and the consequences of an incorrect result to set the interval. Review it after repeated drift or repair. A sensor, indicator, pipework or support change warrants reassessment even if the calendar interval has not expired.
Q6. Can display resolution be used as the acceptance limit?
No. A display increment describes how the result is presented, not its total uncertainty. Reference uncertainty, repeatability, mounting effects and signal noise still matter. Set the acceptance limit from the process requirement and confirm that the test can distinguish a pass from a failure.
Summary
Keep the as-found readings. Confirm that the load path is stable, apply a documented reference through the complete installation, and test the positions and range that matter in service. Accept the result only for the conditions demonstrated; adjustment is one possible outcome of calibration, not its purpose.
