1. Overview of Bridge-Based Measurements
There are many types of bridge-based sensors available, including strain gages, load cells, pressure sensors, and torque sensors. Bridge-based sensors use a passive network of resistors known as a Wheatstone bridge. Most Wheatstone bridge-based sensors use all four legs of the bridge as active sensing elements. However, common strain gages include one, two, or four active sensing elements and are accordingly referred to as quarter-, half-, or full-bridge configurations. In some cases, the sensor does not provide all of the resistors, and the remaining elements needed to complete the Wheatstone bridge must be provided by the measurement device. Because the bridge elements are passive, they require additional signal conditioning in the form of excitation to return a meaningful output.
Figure 1. The Wheatstone bridge is commonly used to measure strain, load, pressure, and torque.
Most often, the measurement device provides excitation as a voltage output connected to two nodes on the bridge sensor as shown in Figure 1 by VEX. The device then measures an input voltage across two nodes of the bridge shown in Figure 1 by VCH. A physical phenomena, such as a change in compression applied to a specimen, slightly changes the resistance of the sensing elements in the Wheatstone bridge. If VEX is held constant, changes in any of the resistors with respect to another will change the output voltage of the sensor. You can see this relationship in Equation 1, which can be rearranged to show that the relative resistance changes, and thus the physical phenomenon, are described by the ratio VCH / VEX.
Wheatstone bridge sensors are very sensitive and typically output no more than a few mV. The output of the sensor varies directly with excitation, so the measurement is typically quoted in mV/V. The numerator refers to the magnitude of the voltage output (mV) of the sensor while the denominator refers to the magnitude of the excitation voltage supplied (V). Keep in mind that mV/V is a unitless ratio of two voltages that describes the relative resistances making up the sensor. Many sensors will specify additional units that allow the user to convert the output into more meaningful engineering units. An example is mV/V/Pa, which is a way to represent the sensor output in mV per volt of excitation per Pascal subjected to the sensor.
2. Common Approach to Measuring Bridge-Based Sensors
To accurately measure the ratiometric output of a bridge-based sensor, you must know both the bridge output voltage and the excitation voltage. You can determine the excitation voltage either by using a precision voltage source or by measuring it. When a precision voltage source is used the excitation is known to within only the accuracy and stability of the source. Additionally, the bridge output measurement must be scaled accordingly in software to get the ratiometric reading. If the bridge is held constant and the excitation voltage changes, the measured bridge output will change as well. This results in an error in the ratiometric measurement unless the change in excitation voltage is measured and used to scale the bridge output voltage. Careful design of the excitation voltage source is necessary to ensure that the contribution of the excitation voltage error is minimized.
A traditional gain stage is commonly used in front of the analog-to-digital converter (ADC) to select the best input range. Gain settings typically range from 1 to 5,000 to compensate for different combinations of sensor output and excitation voltages. Some measurement devices use this same path to measure the excitation voltage to account for the error described above. This method typically uses a switch to temporarily disconnect the sensor input and measure the excitation. This measurement is again used in software to scale the measurement accordingly.
3. NI's Approach to Measuring Bridge-Based Sensors
The NI PXIe-433x bridge input modules, from the SC Express product family, and the NI 9237 bridge input module, from the c series product family,incorporate an analog design that removes the measurement dependence on the accuracy of the excitation voltage. The excitation voltage is continuously sensed by precision circuitry on the modules and used to drive the reference input of the ADC. Using this implementation, the modules return data as a ratio of the bridge output voltage and the excitation voltage. This method continuously and automatically corrects for errors in the accuracy of the excitation voltage.
Consider how the ratiometric approach maps to the concept of bridge sensors described above. Imagine that a sensor is connected to the ADC input and excitation voltage output. Now consider an increase in the excitation voltage. The output of a bridge-based sensor is proportional to the excitation voltage, so if the bridge sensor is held constant, the ratio does not change. The concept of measuring a constant ratiometric output with changes in excitation is the core of the ratiometric approach.
Figure 2. The NI PXIe-433x references the excitation voltage from the ADC.
Figure 2 shows the ratiometric measurement architecture of the NI PXIe-433x bridge input modules. Independent sense lines monitor the excitation output voltage and feed it into the ADC reference input. The remote sense inputs can be connected to sense the excitation voltage directly at the bridge if required to compensate for lead wire resistance desensitization errors.
The NI PXIe-433x bridge input modules incorporate a ratiometric hardware design optimized for bridge-based measurements; thus, the module is not well-suited for absolute voltage measurements. However, the NI PXIe-4339 has a separate voltage input mode to enable a user the flexibility of measuring strain or voltage per channel.
4. Benefits of the Ratiometric Hardware Approach
National Instruments designed the NI PXIe-433x and NI 9237 bridge input modules with a ratiometric design for the key benefit of simplifying the measurement of bridge-based sensors. The resulting architecture removes the dependence on the accuracy of the excitation voltage by continuously sensing the excitation voltage and scaling the measurement directly in hardware. Other benefits of this approach are as follows:
- Increase your measurement stability.
The fact that the mV/V ratio stays constant for changes in the excitation voltage increases measurement stability. Small changes in the environment due to temperature differences typically affect excitation voltages. Because the excitation voltage is sensed and fed back to the ADC, any changes play a smaller role in the accuracy and stability of the measurement. This approach also reduces the excitation design requirements, which improves customer value by saving space and increasing channel density.
- Easily scale to engineering units.
By using the excitation as the source for the ADC reference, the output of the ADC provides measurements that are in units of V/V. This reduces the amount of scaling required to return the appropriate units. Furthermore, the NI-DAQmx driver software include features for bridge-based measurements, including custom channel types for strain, force, pressure, and torque measurements. This allows you to select the appropriate measurement type from the DAQ Assistant or the NI DAQmx-API. Measurements are returned in the physical units that you select from N, lb, Pa, psi, Nm, and so on. With the new channel types, you can choose to enter two-points, a table, or polynomial coefficients depending on the specifications provided by the sensor manufacturer. Combining these features yields a very simple interface to measure bridge-based sensors.
- Acquire measurements at high sampling rates.
Because the excitation is continually fed back to the ADC, there is no additional measurement time required to sample this signal using a switch or another ADC. This allows the NI PXIe-433x bridge input modules to operate at speeds up to 102.4 kS/s per channel.