Force and strain are among the most frequently measured quantities in engineering testing. They decide whether a wing spar is within its design limits, whether a suspension arm will survive a pothole, whether a bolted joint is still clamped, and whether a test rig is applying the load it claims to apply. Two sensors dominate that work: the strain gauge and the load cell. They are often mentioned in the same breath, and for good reason — most load cells are built from strain gauges — but they answer different questions and are specified in different ways.
This guide explains how each one works, the circuit that makes both of them usable, the main types of load cells, when it is better to instrument the component itself, and how to choose between them for a given test. The article is organised into a series of sections, each covering a distinct step in the selection process — from how each sensor works to a practical checklist for choosing between them.
The short answer
- A strain gauge measures strain — the relative deformation of the surface it is bonded to.
- A load cell measures force. It is a complete, calibrated transducer that is placed in the load path.
- Most load cells work by bonding strain gauges to a precisely machined metal element and calibrating the assembly, so that strain in that element can be read directly as force.
In other words, a strain gauge is a sensing element, and a load cell is one of the most common products built with it. The practical choice is usually not between two competing technologies, but between putting a calibrated load cell into the load path and turning the real component into the sensor.
What is a strain gauge?
How a strain gauge works
A strain gauge is a thin, fine grid of electrical conductor — in most gauges a metal foil — on a flexible backing. It is bonded to the surface of a part with a specialised adhesive. When the part is loaded and its surface stretches or compresses, the grid stretches or compresses with it, and its electrical resistance changes in proportion to the strain.
The changes involved are very small. Strain is usually expressed in microstrain (millionths of the original length), and a metal foil gauge has a gauge factor of about 2, meaning its relative change in resistance is roughly twice the strain. On a typical 120 Ω or 350 Ω gauge, a structurally meaningful strain produces a resistance change of a fraction of an ohm. Measuring that reliably, in the presence of temperature changes and electrical noise, is what the rest of the measurement chain is for.
The Wheatstone bridge
Strain gauges are almost always connected in a Wheatstone bridge: four resistive arms arranged so that, with no strain, the output voltage is zero, and any imbalance produces a small voltage proportional to the change in resistance. The bridge is powered by an excitation voltage, and its output is expressed in millivolts per volt of excitation (mV/V).
- Quarter bridge: one active gauge and three completion resistors. Simple and compact, but the most sensitive to temperature effects and lead-wire resistance.
- Half bridge: two active gauges, arranged so that their signals add while temperature effects largely cancel — for example, one gauge in tension and one in compression on opposite faces of a bending beam.
- Full bridge: four active gauges. The highest output and the best compensation for temperature and unwanted load components. It is the configuration used in virtually all strain gauge load cells.
Types of strain gauges
- Linear foil gauges measure strain in one direction, and are the most common choice when the direction of the principal strain is known.
- Rosettes combine two or three grids at defined angles to determine the magnitude and direction of principal strains when they are not known in advance.
- Shear and torque patterns place grids at 45° to measure shear strain, which is how torque on a shaft is measured.
- Semiconductor gauges offer a much higher gauge factor, and therefore a stronger signal, at the cost of greater sensitivity to temperature and non-linearity.
- Weldable and high-temperature gauges are designed for installations where adhesive bonding is not practical or where temperatures exceed what standard gauges tolerate.
Where a gauge cannot be bonded at all — soft materials, very small specimens or a test that runs to fracture — strain can also be measured without contact, optically. See our guide to high-speed cameras, DIC and video extensometry.
What is a load cell?
How a load cell works
A load cell converts force into an electrical signal. Inside, a metal spring element — machined from steel, stainless steel or aluminium alloy — deforms elastically, by a very small amount, under load. Strain gauges are bonded at the points where that element strains most, and wired as a full Wheatstone bridge. Because the geometry of the element is designed and the complete assembly is calibrated against known forces, the bridge output can be read directly in newtons or kilonewtons, typically as a rated output of a few millivolts per volt at full capacity.
The design of the spring element is what gives each type of load cell its capacity, its sensitivity to off-axis loads and its mechanical form. It also explains why a load cell is only accurate when the force passes through it the way it was designed to be loaded.
Types of load cells
| Type | How it is loaded | Typical use |
|---|---|---|
| Bending beam | A beam that bends under load, gauged where bending strain is highest | Low capacities, platform and bench scales, compact test fixtures |
| Shear beam | A beam gauged to measure shear rather than bending, less sensitive to load position | Medium capacities, industrial weighing, test rigs |
| S-type | An S-shaped element loaded along its axis, in tension or compression | Material testing machines, suspended loads, actuator force |
| Tension (in-line) | Threaded ends placed in line with a rod, cable or linkage | Cable and linkage forces, pull tests, hanging loads |
| Compression (column, canister, button) | Load applied to the top of a column or button | High capacities, press forces, restricted installation space |
| Low-profile (pancake) | A flat, stiff element loaded through its centre, often in both directions | Fatigue and structural test rigs, high accuracy with high side-load tolerance |
Not every load cell uses strain gauges. Piezoelectric force sensors excel at highly dynamic and impact forces but are not suited to static measurement, while hydraulic and pneumatic load cells are used where no electrical supply is available or in hazardous areas. For most engineering test work, the strain gauge load cell remains the reference because it measures both static and dynamic forces accurately.
Load cell vs. strain gauge: key differences
| Strain gauge | Load cell | |
|---|---|---|
| What it measures | Strain at a point on a surface | Force along a defined axis |
| What you get | A sensing element that must be installed, wired and, if force is needed, calibrated in place | A complete, calibrated transducer with a stated capacity and accuracy |
| Installation | Bonded to the structure being tested; requires surface preparation and skill | Mounted into the load path; the structure may have to be adapted to receive it |
| Effect on the test article | Negligible — it measures the real part without changing it | Adds a component, compliance and mass to the load path |
| Output | Strain; force only after a calibration of the instrumented part | Force directly, in N or kN |
| Typical use | Structural testing, stress analysis, torque on shafts, instrumented components | Test rigs and actuators, material testing, weighing, force verification |
When to instrument the part instead of adding a load cell
A load cell is the right answer when there is room in the load path for it and the force at that point is what needs to be known. In many real test campaigns there is no such room — the load travels through a suspension arm, a driveshaft, a wing attachment or a bolted joint, and inserting a transducer would change the very behaviour being tested. The alternative is to bond strain gauges to the component itself and, where force or torque is required, calibrate the instrumented part so that it becomes a transducer in its own right.
This approach measures the real part under real conditions. It is how automotive load and effort sensors measure loads in suspension and steering components and torque on driveshafts, propulsion shafts and steering columns.
Instrumented bolts
Bolted joints are a clear example. The clamping force in a bolt determines whether a joint holds, but it cannot be measured with an external load cell without altering the joint. An instrumented bolt solves this by integrating strain gauges — and, in more advanced designs, the strain gauge amplifier as well — inside the bolt head, so the bolt itself reports the force it carries without changing the mechanical behaviour of the assembly. Our instrumented bolts for bolted aerospace structures cover diameters from M4 to M33, detect the border between elastic and plastic behaviour to identify structural limits during testing, and can be deployed on test benches or on site with wired or, on request, wireless connection.
Rotating components
Strain gauges on a rotating shaft, rotor or gearset cannot be cabled directly to a stationary acquisition system. The signals are conditioned and digitised on the rotating part and transmitted across a short gap to a stationary receiver. Our rotating telemetry systems provide wireless torque and strain measurement on engines, turbines, gearboxes and powertrains — see, for example, our automotive powertrain testing case, where digitising close to the strain gauges on the rotor reduces noise.
How to choose: a practical checklist
- Do you need force or strain? Stress analysis and structural validation need strain on the real part. Verifying an applied load needs force.
- Is there room in the load path? If not, instrument the component. If there is, a calibrated load cell is usually faster to deploy and easier to trace.
- What range and overload? Size the load cell so normal loads use a meaningful part of its capacity, while peaks, shocks and misuse stay within its safe overload rating.
- How accurate, and in what sense? Compare non-linearity, hysteresis, repeatability, creep and the effect of temperature on zero and span, not just a single accuracy figure.
- Static or dynamic? Fast transients and impacts need a high natural frequency and suitable acquisition rates; very high-frequency impact forces may call for piezoelectric sensors.
- Which loads should be ignored? Side loads, bending moments and torque can corrupt a force reading. Choose a load cell type and mounting that reject them, or a bridge arrangement that cancels them.
- Where will it live? Temperature range, humidity, ingress protection, vibration and available space often eliminate options faster than the electrical specification.
- How will it be conditioned and recorded? Excitation, amplification, cable length and the data acquisition system are part of the measurement, not an afterthought.
Signal conditioning and data acquisition
Whether it comes from a single gauge or a load cell, a bridge signal of a few millivolts needs a stable excitation voltage, amplification, filtering and digitising before it becomes usable data. Long cable runs add resistance that reduces the effective excitation, which is why many load cells use additional sense wires so that the conditioning electronics can compensate. Conditioning placed close to the sensor, and converted to a digital bus such as CAN, reduces noise and wiring in distributed installations — the role of our signal conditioning modules.
The rest of the chain — sampling rate, synchronisation between channels and recording — is covered in our guide to what a data acquisition system is and how to choose one, and, when the data has to reach the ground in real time, in our guide to telemetry data.
Frequently asked questions
Is a load cell the same as a strain gauge?
No. A strain gauge is a sensing element that measures strain at a point on a surface. A load cell is a complete transducer that measures force. Most load cells contain strain gauges, bonded to a machined metal element and calibrated so that the strain in that element corresponds to a known force.
How does a load cell work?
A force applied to the load cell deforms a metal spring element very slightly. Strain gauges bonded to that element change their electrical resistance with the deformation, and a Wheatstone bridge converts that change into a small voltage. Because the load cell has been calibrated, that voltage corresponds directly to the applied force.
What is the difference between a quarter, half and full bridge?
The number of active strain gauges in the Wheatstone bridge. A quarter bridge uses one active gauge, a half bridge two and a full bridge four. More active gauges give a higher output and better compensation for temperature and unwanted loads, which is why load cells use a full bridge.
Can strain gauges measure torque?
Yes. Gauges bonded at 45 degrees to the axis of a shaft measure the shear strain produced by torque, usually in a full bridge that rejects bending and axial loads. On a rotating shaft, the signals are conditioned on the shaft and transmitted to a stationary receiver through a wireless telemetry link.
How often should a load cell be calibrated?
It depends on how critical the measurement is, how heavily the load cell is used, whether it has been overloaded and what the applicable quality system or test standard requires. Annual calibration is a common starting point, with an additional calibration after any overload, repair or suspect reading.
Specify your force and strain measurement with us
Sensicore supplies strain gauge-based sensors, load and effort measurement, instrumented components, signal conditioning and data acquisition for aerospace, automotive, defence and industrial testing. Explore our precision sensors and specific instrumentation, or contact our engineering team to discuss your measurement.