Hysteresis 101

Accuracy is not a single value you find on a sensor’s specifications data sheet. Instead, it is a composite of several performance factors. Among the most critical and often misunderstood is hysteresis.

What is hysteresis? Referring to force measurement and load cell technologies, technically, hysteresis is the algebraic difference between the sensor’s output at a specific load when approached from an increasing (ascending) direction versus a decreasing (descending) direction.

Imagine you are testing a 1,000 lbf load cell. You measure the output at 500 lbf as you methodically increase the load. Then, you go to 1,000 lbf and bleed the pressure back down to 500 lbf. If the sensor is “perfect,” the two readings at 500 lbf would be identical. Even with the most accurate performing load cells, such as our LowProfiles and Calibration Load Cells, there will still be a gap. This is hysteresis.

Five Hysteresis Basics

#1 – Understanding hysteresis is vital because it represents a physical limit of the sensor’s design. It is a crucial performance specification to review before selecting a load cell for your application. Hysteresis is typically expressed as a percentage of Full Scale (%FS).

#2 – LowProfile Load Cells are engineered for high stiffness and minimal extraneous signal interference. Because they have very little physical deflection compared to a tall column or a flexible S-Beam, the hysteresis is significantly lower—often as low as 0.02% to 0.05% FS.

#3 – Generally, high-quality aluminum exhibits slightly lower hysteresis compared to certain steel load cell designs. Note, hysteresis isn’t just internal to the material.

#4 – Improper mountings, such as not reproducing the torque or surface conditions used during calibration, can introduce mechanical friction, leading to increased hysteresis.

#5 – If your application only involves single-ascending loads (e.g., weighing a tank as it fills), hysteresis is largely irrelevant. However, if you are measuring cycling forces or tension-to-compression transitions, hysteresis becomes a significant source of error to include in your error budget.

Where to Find Hysteresis Data

There are three common places to find the sensor’s hysteresis information.

Product specification data sheets detail hysteresis as a performance characteristic. It is found under MAX ACCURACY, on the Interface data sheets. It is represented as the difference between the output at a given load descending from maximum load and the output at the same load ascending from minimum load. This lists the guaranteed maximum hysteresis the sensor will exhibit under ideal conditions. WATCH:

Calibration certificates contain a load cell’s serial number under performance, along with rated output, SEB output, nonlinearity, and SEB. It is the result of comparing the raw data from ascending versus descending runs. This provides the actual measured difference recorded during the NIST-traceable test for your specific serial number. READ: Calibration Certificates 101

Error budgets, sometimes called uncertainty budgets, for your project must account for the maximum possible deviation in your system. Hysteresis should be included as a line item in your combined uncertainty calculation. It represents the portion of your total measurement error specifically attributed to the direction of loading. READ: How Do You Define A Load Cell Error Budget?

Pro Tip: If your calibration results show a jagged curve or a failure to return to zero, don’t immediately blame the sensor. Check your setup for overshooting during the load application or mechanical interference in the mounting.

The Minor Loop Trap and Hysteresis Sensor Tips

Hysteresis is highly dependent on the loading sequence. A common error during manual testing is overshooting a target load (e.g., aiming for 60% but hitting 80% and then backing off), which causes the sensor to enter a minor hysteresis loop. This results in an output that doesn’t align with the calibration curve, leading to errors that don’t return to zero.

REFERENCE: Learn about this example with details in the performance readings in our Load Cells 301 Guide: Load Cell Characteristics and Applications.

In additon to standard load cells, it is important to evaluate hysteresis for the different sensor model types.

• In a multi-axis sensor, hysteresis can be compounded by crosstalk. This is why multi-axis error budgets are more complex; you have to account for hysteresis across all axes.
• High-quality torque transducers use specialized materials to ensure the zero crossover is as clean as possible. Still, users must be careful not to over-torque, which can permanently shift the hysteresis loop.
• The static error band (SEB) combines non-linearity and hysteresis into a single envelope centered on a best-fit straight line. This provides a more realistic worst-case accuracy figure for complex loading cycles.

Technicians Hysteresis Checklist

The Loading Sequence Test

Perform a calibration cycle where you record data at 50% of the target load twice: once while ascending (increasing from zero) and once while descending (decreasing from full capacity). If the two readings at the same 50% mark differ beyond your specification, hysteresis is present.

  The Return to Zero Check

After applying a full-scale load and then removing it completely, does the sensor return to its original zero balance? A significant residual offset that is not caused by creep over a long duration is a classic indicator that the sensor is holding a memory of the previous load.

  The Overshoot Observation

During your test, did the operator or hydraulic system accidentally exceed the target load and then back off to the correct value? If so, you have entered a minor hysteresis loop. If your data point looks like an outlier compared to the rest of the curve, this is the likely culprit.

  The Mounting Torque Audit

Check the mounting bolts. Are they torqued exactly to the manufacturer’s specifications? Under- or over-tightened bolts can cause the mounting surfaces to slip or bind microscopically during loading, introducing mechanical hysteresis that appears as sensor failure.

  The Friction and Interference Visual

Inspect the entire load string. Is any part of the live load touching a non-live surface, such as a safety shroud, a cable tied too tightly, or a bent blast shield? Even minor mechanical rubbing creates friction that prevents the sensor from returning to its natural state, artificially inflating hysteresis values.

By optimizing your test protocol to avoid overshooting and ensuring your mounting conditions match your calibration, you can minimize its impact and ensure the highest integrity for your data. We recommend reviewing our Load Cell Basics webinar to better understand the essential accuracy characteristics of load cells. For technical topics, be sure to read our Contributing Factors To Load Cell Accuracy white paper.