Simple Steps to Selecting the Right Accelerometer

Simple Steps to Selecting the Right Accelerometer

Sensors Insight by Bruce Lent

Looking for an accelerometer? Faced with choices of technology, shape, size, and sensing range, novices as well as more experienced users can be intimidated when examining an accelerometer manufacturer's catalog or Web site. The path described in this article will help you find your way through the thicket of options to choose the optimum accelerometer for your application.


Technology Selection

The first step in the selection process is to determine the type of measurement to be made. There are three popular technologies used for acceleration measurements.

Piezoelectric (PE) accelerometers are the most widely used accelerometers for test and measurement applications. These devices offer a very wide measurement frequency range (a few Hz to 30 kHz) and are available in a range of sensitivities, weights, sizes, and shapes. PE accelerometers are available with a charge or voltage (IEPE) output, discussed later in this article. They are appropriate for both shock and vibration measurements.

Piezoresistive (PR) accelerometers generally have low sensitivity making them desirable for shock measurements and less useful for vibration measurements. They are also used extensively in transportation crash tests. PR accelerometers generally have a wide bandwidth (from a few hundred Hz to >130 kHz) and the frequency response goes down to 0 Hz (often called "DC responding") or steady state, so they can measure long-duration transients.

Variable capacitance (VC) is among the newer accelerometer technologies. Like piezoresistive accelerometers, VC accelerometers are DC responding. They have high sensitivities, a narrow bandwidth (from 15 Hz to 3000 Hz), and outstanding temperature stability. Thermal zero and sensitivity shifts can be as low as 1.5% over a temperature range of 180°C. These devices are suited for measuring low-frequency vibration, motion, and steady-state acceleration.


Type of Measurement

First, we'll describe the basic measurement types, providing more detail later. For the purposes of this article, we will divide acceleration measurements into the following categories:

Vibration—An object is said to vibrate when it executes an oscillatory motion about a position of equilibrium. Vibration is found in the transportation, aerospace, and industrial environments as well as when simulated by a shaker system.

Shock—A sudden transient excitation of a structure that generally excites the structure's resonances. A shock pulse can be produced from an explosion, a hammer striking an object, or a vehicle crash.

Motion—Motion is a slow-moving event lasting from <1 s to several minutes, such as the movement of a robotic arm or an automotive suspension.

Seismic—This is more like a slow-motion or a low-frequency vibration. This measurement usually requires a specialized, low-noise, high-resolution accelerometer. Seismic accelerometers are used to measure the motion of bridges, floors, and earthquakes.


General Considerations

Before we discuss the technologies and applications in greater detail, here are a few general considerations.

Frequency response is the accelerometer's electrical output vs. a mechanical excitation over a frequency range with a fixed amplitude and it is an important parameter when considering any accelerometer. The frequency range will usually be determined by the test specifications or by the user. It is usually specified within ±5% of the reference frequency (usually 100 Hz). Many devices will have the specifications extended to ±1 dB and in some cases ±3dB. The preceding limits express the accuracy that can be expected over a given frequency range. Most data sheets will have a typical frequency response curve to assist the user. The curves illustrate how the accelerometer's accuracy varies over a specified frequency range.

Another consideration is the number of axes to be measured. Accelerometers are available in single-axis and triaxial (3-axis) versions (Figure 1). An alternative approach to making a three-axis measurement is to mount three accelerometers on a triaxial mounting block. Both methods allow for the measurement of three orthogonal axes simultaneously.

Fig. 1: A miniature triaxial accelerometer incorporates three accelerometers in a single package to simultaneously measure acceleration in three orthogonal axes.
Fig. 1: A miniature triaxial accelerometer incorporates three accelerometers in a
single package to simultaneously measure acceleration in three orthogonal axes.



Piezoelectric accelerometers are the first choice for most vibration measurements since they have a wide frequency response, good sensitivity and resolution, and are easy to install. There are two types of piezoelectric accelerometers: the basic charge-mode accelerometer and the voltage-mode Internal Electronic Piezoelectric (IEPE) accelerometer.

In recent years, the IEPE type has become the most commonly used accelerometer type because of its ease of use. IEPE sensors are often sold under different trademarked names, but most comply with a pseudo industry standard and are interchangeable between brand names. (Author's note: Be sure that the power source current and voltage are compatible with the accelerometer you've selected to achieve optimal performance and to avoid possible damage.) Basically, an IEPE accelerometer has a charge amplifier built into the accelerometer. As a result, the sensor requires no external charge amplifier and uses ordinary, low-cost cable. The accelerometer does require a constant current power source and many DA systems have built-in power sources. For a known vibration range and an operating temperature that lies within the range of –55C°C to 125°C consider using an IEPE device. Note that high-temperature versions of some models have a maximum operating temperature of 175°C.

The advantages of charge-mode piezoelectric accelerometers include high-temperature operation and an extremely wide amplitude range, which is largely determined by the charge amplifier setting. An IEPE accelerometer has a fixed amplitude range. A typical charge-mode accelerometer will have an operating temperature range of –55°C to 288°C. Special-purpose accelerometers are available for extreme environments as low as –269°C to as high as 760°C. Special radiation-hardened charge-mode accelerometers are available for use in a nuclear environment.

Unlike the IEPE accelerometer, the charge-mode accelerometer requires the use of a special low-noise cable, which is expensive when compared to the standard commercial coaxial cable. A charge amplifier or an in-line charge converter is also required for operation. Charge-mode accelerometers are preferred for high-temperature operation (above 175°C) or in cases where the maximum acceleration is unknown.

In instances where vibration measurements at very low frequencies are required, consider choosing a VC accelerometer (Figure 2). VC accelerometers have a frequency response from 0 Hz to 1 kHz, depending on the sensitivity required. When making very low frequency measurements, a VC accelerometer with a frequency range from 0–15 Hz will provide a sensitivity of 1 V/g. VC accelerometers are useful on electrohydraulic shakers, to make flutter measurements, and for many transportation applications, such as testing automotive and suspension systems and making railroad ride and sway measurements.

Fig. 2: Variable capacitance accelerometer shown with a popular mounting configuration.
Fig. 2: Variable capacitance accelerometer shown with a popular mounting configuration.



Two technologies are available for shock measurements and, depending on the shock levels and the final data required, you can choose from a variety of accelerometers. It is important to know the expected shock level since this will determine the type of accelerometer to be used. Here is a rough guide to assist the reader in choosing the proper accelerometer.

Table 1

For low-level shock measurements, a general-purpose accelerometer will usually do the job. The accelerometer will need a linear range of at least 500 g and a shock survivability rating of 500 g. An IEPE type is usually preferred because they are less susceptible to producing erroneous results from cable motion. Use an amplifier with a low-pass filter to attenuate the accelerometer resonance.

For automotive crash testing, a rather specialized area of shock testing, piezoresistive accelerometers are usually used (Figure 3).

Fig. 3: An automotive crash test accelerometer shown in an industry standard package.
Fig. 3: An automotive crash test accelerometer shown in an industry standard package.

For far-field shock measurements, a special shear-mode accelerometer with a built-in electronic filter is often adequate. These are usually lightweight IEPE types with solder connections. The electronic filter attenuates the resonance frequency of the accelerometer to prevent overloading of the DA equipment.

Near-field measurements are often more than 20,000 g. Here the choice of accelerometers is dependent on the type of test being conducted. Specialized accelerometers of either piezoelectric (charge-mode and IEPE) or piezoresistive may be appropriate (Figure 4). Typically, an IEPE with characteristics like the far-field accelerometer is appropriate, but with the addition of an internal mechanical filter. The mechanical filter will ensure the survivability of the accelerometer and will generally eliminate zeroshift.

Fig. 4: An example of a near-field shock accelerometer with built-in mechanical filter and rugged 1/4-28 mounting stud.
 Fig. 4: An example of a near-field shock accelerometer with built-in mechanical filter and rugged 1/4-28 mounting stud. 

A detailed discussion of zeroshift is beyond the scope of this article, but, in general terms, the zeroshift phenomenon appears when the time history doesn't return to the zero-acceleration level following the shock event. This shifting results in distortion of data when performing integration. Zeroshift is rare when piezoresistive accelerometers are used.

As is the case with vibration, the frequency response is an important parameter for shock. In general, a shock accelerometer should have a wide frequency response range (10 kHz is typical), depending on what is being tested.


Motion, Constant Acceleration, and Low-Frequency Vibration

VC accelerometers should be considered for applications within this category. This technology allows for the measurement of low-level, low frequency vibration with a high output level. They also provide a high degree of stability over a broad temperature range.

When a VC accelerometer is placed in a position where the sensitive axis is parallel to the earth's gravity, an output equal to 1g will be produced. This phenomenon is often referred to as "DC responding." Because of this characteristic, VC accelerometers are very useful for measuring centrifugal force or for measuring acceleration and deceleration of devices such as elevators.

In the realm of vibration testing, VC accelerometers are used in applications where low-frequency events are to be studied and where preservation of phase data is important. VC accelerometers have found their niche around aircraft flutter testing. Their low-frequency characteristics make VC accelerometers ideal for ride quality measurements in automobiles, trucks, and railroad equipment. A wideband frequency response is not a characteristic of VC devices.


The Environment

Once the technology has been selected and the test type determined, there are several other factors to be considered. As a starting point, considering the environment in which the sensor will operate. Environmental characteristics include temperature, maximum acceleration levels, and humidity.

The table in Figure 5 shows typical values to assist with temperature selection:

Fig. 5: Accelerometer technologies based on operating temperature ranges.
Fig. 5: Accelerometer technologies based on operating temperature ranges.

The specified g range is often confusing to the new accelerometer user since this parameter appears twice in the specifications. The actual usable range of the accelerometer is found in the dynamic specifications. For example, an IEPE accelerometer might have a range of 500 g and, under the environmental characteristics, the device has a shock limit of 1000g and a shock limit of 2000g. The 500g is the maximum range of linear operation of the accelerometer. The parameters specified in the environmental section are indicative of the maximum survivable shock and/or sine acceleration levels.

In the case of charge-mode piezoelectric devices, a range is not specified under the dynamic characteristics since it is largely determined by the charge amplifier. The user should refer to the amplitude linearity specifications, in the dynamic characteristics section of the data sheet. As above, the maximum range specified in the environmental characteristics section is a maximum survivability figure.

The humidity specification is usually given as "Hermetic," "Epoxy Seal," or "Environmental Seal." Most of these seals will withstand high levels of moisture. If the accelerometer is being used in the space environment, underwater, or with very long exposure to excessive humidity, a hermetic seal is recommended. It should be noted that continuous temperature cycling can make an epoxy seal fail.

If accelerometers are designed to operate within a nuclear radiation environment, the data sheets will so indicate. Magnetic susceptibility is seldom specified since it is usually not a problem with newer accelerometers. Nonmagnetic materials are used in modern accelerometers, thus reducing this problem.

If the accelerometer is going to be mounted on a highly flexible surface, the base strain specification becomes important. A flexile surface tends to bend, inducing strain on the accelerometer's base. The resulting strain can appear as vibration in the accelerometer's signal, distorting the output. As a general guide, avoid using compression-type accelerometers on flexible surfaces.


Accelerometer Weight

When an accelerometer is attached to the test article, the measured acceleration will be altered. These effects can be reduced to an insignificant amount by being mindful of the accelerometer's weight (Figure 6). As a rule-of-thumb, the weight of the accelerometer should be no greater than 10% of the weight of the test article.

Fig. 6: On the left is a popular side-connector accelerometer that weighs 7.8 g and is used on heavy test articles (A). The miniature accelerometer (B) on the right weighs 0.5 g and can be mounted on lightweight structures and PC boards.
Fig. 6: On the left is a popular side-connector accelerometer that weighs 7.8 g and is used on heavy test articles (A). The miniature accelerometer (B) on the right weighs 0.5 g and can be mounted on lightweight structures and PC boards.



There are a few ways to mount an accelerometer to the unit under test (UUT), and methods include everything from permanent mounting to temporary methods. Here are a few of the most common mounting methods.

The best mounting method uses a threaded stud or screw. Stud/screw mounting provides the best transmissibility at high frequencies since the accelerometer is virtually fused to the mounting surface. High-frequency response can be enhanced by the application of light oil between the accelerometer and the UUT. If this method of mounting is desired, accelerometers should be purchased that are designed for stud and/or screw mounting.

Adhesive mounting is often required, especially on small surfaces and PC boards. The preferred mounting adhesive is a cyanoacrylate because it can be easily removed (with the proper removal techniques). Many accelerometers are specifically designed for adhesive mounting and this fact will be noted on the data sheet. A stud-mount accelerometer may be mounted using an adhesive, but a cementing stud should be used to prevent the adhesive from damaging the accelerometer's threads.


Ground Isolation

Ground isolation becomes important when the test article's surface is conductive and at ground potential. A difference in ground voltage levels between the electronic instrumentation and the accelerometer may cause a ground loop resulting in erroneous data.

Accelerometers are available with ground isolation or with the ground connected to the accelerometer's case. Accelerometers with ground isolation usually have an isolated mounting base and, where applicable, an isolated mounting screw. In some cases, the entire accelerometer case is ground isolated.


Sensitivity and Resolution

When either a low-level signal and/or a wide dynamic range is required, the accelerometer's resolution and sensitivity become important. An accelerometer converts mechanical energy into an electrical signal (the output). The output is expressed in terms of millivolts per g (mV/g), or, in the case of a charge-mode accelerometer, the output is expressed in terms of pico-Coulombs per g (pC/g). Accelerometers are offered in a range of sensitivities and the optimum sensitivity is dependent on the level of the signal to be measured e.g., in the case of a high g shock test, low sensitivity is desirable.

In the case of low-level signals, the best approach is to use an accelerometer of high sensitivity to provide an output signal well above the amplifier's noise level. For example, if the expected vibration level is 0.1g and the accelerometer has a sensitivity of 10 mV/g, then the voltage level of the signal would be 1 mV, and a higher sensitivity accelerometer may be desirable.

Resolution is related to the accelerometer's minimum discernable signal. This parameter is based on the noise floor of the accelerometer (and in the case of an IEPE type, the internal electronics) and is expressed in terms of grms.


Other Considerations

The above information will help the potential user make a preliminary decision as to which accelerometers can potentially perform the measurement task. However, there are other—equally important—parameters that should be discussed with potential suppliers. These important items may include:

  • Signal conditioning and powering
  • Transverse sensitivity
  • Temperature response
  • Cable types

Once the questions posed in this article have been answered, further discussion with the manufacturer is recommended.


About the author

Bruce Lent is a Senior Applications Engineer at Meggitt PLC.