Strain gages are fundamental sensing devices that function as the building blocks of many other types of transducers, including pressure, load, and torque, used extensively in structural test and monitoring applications. Even though strain gages are very common, they are one of the most difficult types of sensors from which to condition and acquire reliable data. Strain gage measurements operate by sensing minute changes in length of a metal foil due to stress across a surface that is often smaller than 5 mm2.
Several factors can affect the measurement performance of your strain gage. One set of factors is related to signal conditioning, including the Wheatstone bridge, and the excitation source. Another set is related to properties of the foil gage, including self-heating and transverse sensitivity.
Figure 1. Basic Wheatstone Bridge Circuit Diagram
Because of the complexity of foil gage measurements, there are countless factors to consider. Some have negligible effects on measurement performance, and others have effects that are much more pronounced. Here is a core set of considerations that address the major causes for inaccuracies in your strain measurements. These five considerations can greatly improve the quality of your strain measurements.
1. Bridge Type and Configuration
Wheatstone bridges are the most commonly used method for measuring the electrical resistance of strain gages. These incorporate four resistive elements, where one or more can be active strain sensing elements. The number of active elements dictates its configuration: either quarter- (one active), half- (two active), or full-bridge (four active). There are also different types, depending on the position of the gages and type of force being measured. To select the most appropriate bridge configuration for a specific application, you should consider several parameters. Table 1 can help simplify this process.
Table 1. Bridge Configuration Comparison of the Top Characteristics
Type of strain. The two most common types of strain that can be measured are axial (compression/tension) and bending. Some bridge configurations will reject either axial or bending strain.
Figure 2. Axial and Bending Strain
Temperature compensation. Ideally, strain gage resistance should change in response to strain only. However, a strain gage’s resistivity and sensitivity also change with temperature, inducing measurement errors. Temperature-compensated bridge configurations are more immune to temperature effects.
Transverse sensitivity. This is also known as the Poisson Effect. An ideal strain gage should respond to strain only along its intended longitudinal axis. However, as the strain gage elongates due to a tensile force, it also contracts in the transverse direction, which affects the measurement. You can compensate for this by measuring the transverse force with a Poisson gage and applying the appropriate equation.
Measurement sensitivity. For the same strain gage, changing the bridge configuration can improve its sensitivity to strain. For example, the Full-Bridge Type I configuration is four times more sensitive than the Quarter-Bridge Type I configuration. However, Full-Bridge Type I requires three more gages than Quarter-Bridge Type I and requires access to both sides of the gaged structure.
Installation. Installing strain gages can take a significant amount of time and resources, and the amount varies greatly depending on the bridge configuration. The number of bonded gages, number of wires, and mounting location all can affect the level of effort required for installation. Certain bridge configurations even require gage installation on opposite sides of a structure, which can be difficult or even impossible. Quarter-Bridge Type 1 is the simplest, requiring only one gage installation and two or three wires.
Figure 3. Structures with Gages Installed on One Side and on Opposite Sides
As a side note, the NI-DAQmx driver software takes care of calculations and equations related to temperature compensation, Poisson ratio, and bridge configuration type, and returns microstrain (µe) physical units. NI hardware also provides bridge completion using high-precision resistors on the measurement device and/or connectivity accessory.
2. Lead Wire and Bridge Arm Resistance
Lead wire resistance can cause a reduction in sensitivity because of long wires and small-gauge wires, which have greater resistance than the bridge completion wiring within the measurement system. In addition, for half- and full-bridges, long lead wires and small-gauge wires can deliver a lower excitation voltage across the sensing element than originally intended.
Two ways to correct these errors are to use remote sensing and perform shunt calibration.
For half-bridge and full-bridge configurations, you can connect remote-sense wires to the points where the excitation voltage wires connect to the bridge circuit, as seen in Figure 4.
Figure 4. Remote-Sense Wire Connections with an NI 9237 Simultaneous Bridge Module
NI measurement devices such as the NI 9237 and PXIe-4330 can use remote-sense leads to measure the actual excitation voltage seen at the bridge. The actual bridge excitation voltage could be smaller than the voltage at the EX+ and EX– leads because of the added lead wire resistance. Without remote sensing of the actual bridge voltage, the resulting gain error is Rlead/Rbridge for half-bridge sensors, and 2*Rlead/Rbridge for full-bridge sensors.
For quarter-bridge configurations, a reduction in sensitivity could be induced because of the resistance of the leads in the sensing arm of the bridge. You can use shunt calibration to compensate for this reduction in sensitivity. The NI 9237 shunt calibration circuitry, for example, consists of a precision resistor and a software-controlled switch. The shunt calibration procedure simulates the input of strain by changing the resistance of the sensing arm in the bridge by some known amount. This is accomplished by shunting, or connecting, a large resistor of known value in parallel to one arm of the bridge, creating a known change in resistance. The output of the bridge is then measured and compared to the expected voltage value to correct gain errors in the entire measurement path.
NI measurement devices for strain gages all offer shunt calibration circuitry, consisting of a precision resistor and a software-controlled switch. It should be noted that offset nulling should always be performed prior to shunt calibration.
When you install a strain gage and connect it to the Wheatstone bridge for the first time, it is very unlikely that you will read exactly zero volts when no strain is applied. Strain gage imperfections, lead wire resistance, and a pre-strained installation condition will generate some nonzero initial voltage offset. There are a few different ways of handling this initial offset voltage.
One method is to compensate for the initial voltage in software. With this method, you take an initial measurement before you apply input. You then use this initial voltage in the strain equations listed in the Strain Gage Configuration Types tutorial. This method is simple, fast, and requires no manual adjustments. The disadvantage of the software compensation with traditional measurement systems is a loss in effective measurement range due to large offsets. You can solve this through measurement instrumentation that has a wide measurement range. For example, the NI 9237 provides a wide 25 mV/V input range with highly precise 24-bit analog-to-digital converters (ADCs).
4. Excitation Level
The excitation level impacts your strain gage measurements in two ways: the signal-to-noise ratio and gage self-heating. In an ideal world, high excitation voltage levels are preferred to provide a strong signal-to-noise ratio that you can easily measure, especially in noisy environments or in cases where long, noise-susceptible lead wires are used. However, because foil gages are essentially resistive electrical devices, higher excitation levels will cause self-heating, which introduces multiple negative effects.
Self-heating changes a strain gage’s resistivity and sensitivity and the adhesive’s ability to transfer strain, and also introduces thermocouple effects between the lead wires and the foil gage. This effect is especially detrimental if the measured structure does not provide good heat dissipation (such as plastic), and in cases where the foil gage has a very small surface area. You can reduce self-heating by either selecting a strain gage with a bigger surface area for better heat dissipation, or reducing the excitation level. In noisy environments, you can still use low excitation levels by properly shielding the lead wires and placing the measurement device close to the sensors. National Instruments provides a wide selection of form factors, including distributed form factors, which offer maximum flexibility in the placement of the measurement system. NI also offers user-selectable excitation voltages. Ultimately, you must strike a balance when choosing the appropriate excitation level.
5. Excitation Stability
The accuracy of a bridge-based measurement is directly proportional to the stability of the excitation source. Changes in the excitation source will cause changes in the measured output of the bridge. As a result, small excitation source fluctuations will translate to a misrepresentation in strain.
There are two methods to get around unstable and inaccurate excitation sources: one is by measuring the voltage actually supplied by the source to compensate for fluctuations; the other is by referencing the measurement performed by the ADC against the excitation source. The first method requires more power and adds more cost. You implement the latter ratiometric approach in the measurement hardware itself, allowing the ADC to track any fluctuations in the excitation source without any software compensation or additional measurements. The NI 9237 and PXIe-4330 modules provide this functionality.
Understanding these fundamental concepts and considerations can save you time and resources during the planning, purchase, design, and development phases of your structural test and monitoring systems. Ultimately, these considerations can help take the stress out of your strain measurements.
Nathan Yang is product marketing manager for structural test and monitoring at National Instruments. Nathan is responsible for the product management and marketing of data acquisition devices for stress and strain measurements. He holds a bachelor’s degree in electrical engineering from McGill University in Montreal, Canada.