Before sending a suspect load cell for calibration, run a repeatability test. It takes ten minutes, needs only a known load and a way to read the indicator, and it separates a failed cell from a failed installation.
The procedure
With the system unloaded, note the zero reading. Apply a load of about 80% of working capacity, hold for 30 seconds, note the reading, remove the load, and note zero again. Repeat three times without adjusting anything. Record ambient temperature at the start and end.
Reading the result
If the loaded reading repeats within the datasheet repeatability and zero returns each time, the cell is behaving. If the loaded reading drifts upward across cycles, check for a loose mechanical connection that is settling. If zero walks in one direction, suspect thermal drift from self-heating, moisture, or a cable problem. A large gap between loading and unloading readings at the same load is hysteresis, usually mechanical.
What it does not tell you
Repeatability says nothing about span accuracy, so a cell can pass this test and still read 2% high. Use it as a health check, not as a calibration. For trade use, the formal calibration still has to be done at the required interval.
Kevin Huang