Simcenter Testing Solutions All About Accelerometers

The sensitivity, range, and frequency response need to be selected appropriately based on the measurements to be performed.

2.1 Sensitivity

The sensitivity of an accelerometer describes how many volts will be output per unit of acceleration.  The higher the sensitivity value, the more sensitive the accelerometer. Sensitivities are typically given in quantities of voltage/acceleration, such as 10 milliVolt per g (mV/g) of acceleration. This means that for every 10 mV the sensor outputs, it is reading an acceleration of 1 g. A 100 mV/g accelerometer has a higher sensitivity than a 10 mV/g accelerometer. Some typical accelerometers and their corresponding sensitivities are shown in Figure 3:
 

Selecting the sensitivity is a trade-off between getting a good signal to digitize without overloading.  Higher sensitivity accelerometers output a larger voltage.  A larger voltage signal is easier to digitize for analysis on a PC.  However, if the voltage is too high, the signal will be clipped by the data acquisition system's fixed range (typically 10 Volts).  This causes an overload which is not desirable.

For example, different sensitivity accelerometers might be considered modal testing versus operational vibration testing. A typical use of 100 mV/g accelerometers is low excitation vibration measurements such as modal analysis. Operational vibration conditions have higher accelerations than a typical modal test.  For operational measurements, typically 10 mV/g accelerometers can be used. Using a sensitive 100 mV/g accelerometer in a operational vibration test can easily lead to overloads, while using a 10 mV/g accelerometer for a modal test may lead to poor coherence and noisy FRF measurements.

More information on overloads in the knowledge articles: Overloads and Gain, Range, and Quantization.

2.2 Range

Another important aspect in choosing an accelerometer is the maximum range. Some accelerometers can only read up to a certain acceleration level, such as ±500 g peak. Anything outside of this range will not be registered or could register as an overload. 

Converting between sensitivity and voltage can be useful in understanding the expected levels from accelerometers. The sensitivity times the measurement range equals the maximum voltage output for a given accelerometer as shown in Figure 4:
 

It is helpful if the maximum voltage that an accelerometer can output is lower than the input voltage range of the measurement system. For example, on a typical Siemens SCADAS data acquisition system, the maximum range is 10 Volts.

More information about matching an accelerometer output to a measurement system input in the knowledge article: Gain, Range, and Quantization.

2.3 Frequency Response

In general, accelerometers need to have a linear response over the frequency range being measured. This ensures that a single sensitivity value can be used to convert the output voltage into acceleration. This is not possible if the sensitivity of the accelerometer changes with frequency.

At high frequencies it is possible that the output of the accelerometer varies due to a natural frequency of the accelerometer itself. At the natural frequency of the accelerometer, more voltage is output for the same acceleration.

In the Figure 5 below, there are three different sensors that show different accelerometer linearities.
 

For the sensors in Figure 5:

When choosing an accelerometer, it is important to know the highest maximum frequency desired for measurement. This will be important in selecting an accelerometer from a manufacturer. 

The maximum useable frequency range of the accelerometer is not limited by the resonance frequency peak in the accelerometer. There is a flat range frequency as shown in Figure 6:
 

Figure 6: Example of useable bandwidth of an accelerometer.
 

The flat usable bandwidth of the accelerometer is what a user should look at when selecting a sensor. The high frequency of the range must be sufficiently below the peak frequency of the accelerometer resonance.  The low frequency is often determined by a high pass coupling filter used with many IEPE/ICP accelerometers.

3. Accelerometer Types

There are three main types of accelerometer technologies: Piezoelectric (PE/IEPE), Piezoresistive (PR), and Variable Capacitance (VC).

3.1 Piezoelectric

Piezoelectric accelerometers are accelerometers that utilize the piezoelectric effect to measure acceleration. The word “Piezo” comes from Greek meaning “pressure” or “to squeeze”.

When certain types of crystal materials such as quartz or ceramics are stressed or squeezed, the crystal will output a charge proportional to the stress. (Figure 7).

In Figure 8, the forced rearrangement of atoms in a quartz crystal made of silicon dioxide will affect how the atoms are distributed and change the charge
 

Figure 8: The atoms of a quartz crystal undergoing stress and displaying the piezoelectric effect.
 

• Piezoelectric (PE) or Charge Accelerometers
• Integrated Electronics Piezoelectric (IEPE, integrated electronics piezoelectric).

IEPE is also sometimes known as ICP™ (integrated circuit piezoelectric) or as CCP (constant current piezoelectric).

3.1.1 Charge (PE) 

Piezoelectric (PE) sensors, sometimes referred to as charge accelerometers, were the first type of piezoelectric accelerometer. These accelerometers do not have any integrated electronics and the only signal that is measured is the tiny amount of electric charge produced by the PE sensor, typically measured in picoCoulombs (pC):

• High Temperatures: Because of the lack of internal electronics, these sensors are ideal for high or low temperature environmental testing, such as exhaust measurements, or areas with freezing temperatures.
• Triboelectric Effect: Since the signal coming from the piezoelectric is small, it is easily corruptible and is sensitive to the Triboelectric effect. The triboelectric effect is noise in a signal cable that occurs from the rubbing of two dissimilar metals. The cable materials, if not properly secured, can move around freely and will rub against each other creating noise (charge in the cable) that can increase the signal to noise ratio to levels that would not allow the sensor to measure properly. As a counter measure, "low noise" cables should be applied with PE sensors. These cables contain conducting carbon particles that attenuate the triboelectric effect. Cables also need to be secured to prevent vibration motion from exacerbating this effect (Figure 11). 

• With External Signal Conditioner: To power the sensor, there is an external voltage amplifier that provides a constant current excitation that powers the accelerometer. Two standard coaxial cables can be used, and a standard data acquisition system can be used to measure data.
• Without External Signal Conditioner: To reduce the need for two cables, the second configuration shows a SCADAS measurement system with a single cable connection to the accelerometer. IEPE sensors are compatible with a multitude of cards, include V8, VB8, and VC8. Each of these cards can supply the sensor with the constant current excitation needed to power the accelerometer. 

Some other key points about IEPE/ICP/CCP accelerometers:

• Dynamic Only, No Static: Due to the need for external power, these sensors cannot measure static acceleration (e.g. gravity). To separate the useful signal from the DC-excitation,  a high pass filter is automatically activated. It specifically uses a low capacitance quartz with low impedance sensors, making it simpler for measurement.
• Regular Cable: Because of the low impedance sensor, this type of sensor does not require a special noise cable and can be used with a single standard coaxial cable. Due to not having special cable requirements and their simple setup, IEPE sensors are more common in the field than PE charge accelerometers.
• High Temperature: Since IEPE accelerometers contain internal electronics, they are temperature limited and not ideal for high temperature testing. Check the temperature specification if the test involves a hot environment (exhaust system, heat exchanger, etc).
• Transducer Electronic Data Sheets (TEDS): IEPE sensors have one nice additional advantage, they are compatible with transducer electronic datasheet (TEDS) technology. TEDS allow for the serial number, manufacturer, model number, calibration value and calibration expiration date to be read off individual accelerometers. This can make testing setup even quicker and have less room for entry errors.

IEPE/ICP/CCP sensors are compatible with SCADAS V8, VC8, and VB8 cards. 

3.1.3 Pros and Cons of Piezoelectric Sensors

A quick summary of the differences between the two types of piezoelectric technologies can be found in Figure 15.
 
The main benefit of IEPE setups is the simpler and less expensive set up. Less cabling is typically needed, and no additional amplifier is required. The benefit of a PE sensor is the ability for testing in extreme environments. 

3.2 Piezoresistive

Piezoresistive (PR) sensors work similarly to their piezoelectric counterparts. These types of sensors will produce a change in resistance when a mechanical strain is applied to them.

Piezoresistive accelerometers will measure a 1 g offset due to gravity while a piezoelectric accelerometer does not. This is useful in applications where static acceleration needs to be measured (example: the number of g's felt during a ride and handling maneuver) in addition to dynamic acceleration. PR sensors are DC coupled so they are sometimes referred to as DC accelerometers (Figure 16).
 
Many different designs exist for DC accelerometers, but the basics are the same. A proof mass with a known value is attached to a cantilever beam. Connected to the beam are piezoresistive elements or strain gauges. As the mass experiences acceleration, the beam will bend and this mechanical strain will change the resistance value of the strain gauge. The voltage value can then be correlated to an acceleration since the bending strain is proportional to the mass’s acceleration. 

Since DC accelerometers measure the change in resistance, a constant DC excitation is required. The flat DC voltage signal will change based on the resistance change. The piezoresistive elements used in these sensors are typically comprised of a metal strain gauge or silicone Wheatstone bridge. Due to their simple setup, these accelerometers can be made much smaller than PE and IEPE sensors. These sensors are also good for use in crash and shock applications. To use with a SCADAS hardware, a VB8 module is required to provide the DC coupling and power needed. 

More information in the knowledge article: AC versus DC Coupling - What's the difference?.

3.3 Variable Capacitor

A variable capacitor is one that measures a change in capacitance instead of a change in resistance or voltage output. It has a minimum of two parallel plate capacitors that work in differential mode. These work by altering the peak voltage generated by an oscillator when the structure is undergoing acceleration. The simplest of designs involve a known mass that is connected to a moveable electrode (Figure 17).
 
As the mass experiences vibration, the electrode will move up and down between two other fixed electrodes, changing the capacitance of each one. The channels that measure capacitance are set to differential mode to find the difference in capacitance C₁ and C₂. It is called a variable capacitor since the position can change between the two fixed electrodes, making the overall capacitance a variable. These sensors will also need external voltage to power the internal electronics.

3.4 MEMS

Most VC accelerometers are MEMS accelerometers. MEMS stands for micro-electromechanical systems. They are typically made from silicon and are micro-fabricated, meaning these accelerometers are very thin and tiny. MEMS accelerometers are typically used in electronic applications such as a cellphone. In more complex designs, such as Figure 18, a known mass or “proof mass” is created as “fingers”.
 
These fingers are connected in such a way, that their side-to-side movement is not restricted. Fixed electrode fingers are situated in the middle which do not move. As the MEMS accelerometer experiences acceleration, the proof mass fingers will move back and forth, changing the capacitance between the proof mass and the fixed electrodes. Because these accelerometers are also DC coupled, they can measure static acceleration at zero Hz, (gravity). MEMS accelerometers are not typically used in engineering research and development.  They are typically found in industrial or mass produced applications like embedded vibration sensors, cell phones, etc.

4. Accelerometer Setup

Selecting an accelerometer type based on desired frequency range, dynamic versus static, temperature, etc is only part of what needs to be considered when performing a vibration measurement campaign.  Mounting the accelerometer on the test article and cabling are also important.

4.1 Mounting 

There are many different considerations to be made when mounting an accelerometer on a test object.  These include positioning, adhesive, and surface contact.

Some engineers choose to adhere an accelerometer with beeswax since it does not leave permanent damage to a test article. Others use a quick curing super glue, magnet, or even thread the accelerometer into a metal stud affixed to the structure. Each of these mounting techniques has pros and cons. Consulting the chart below (Figure 19), one can see that the stronger the attachment between the accelerometer and the test article, the higher in frequency an accelerometer can measure.
 
Silicone grease improves high frequency measurement since it reduces the frictional damping that occurs between the accelerometer and the test surface. For some test articles, drilling into it to mount an accelerometer is not possible. In the field, sometimes engineers will use a thin piece of polyimide tape on the surface of the test article before mounting an accelerometer with glue. This prevents the test article from getting covered in super glue where surface quality is important, such as aerospace applications. 

In general, a smooth and flat mounting surface is the best for accelerometers, but this may not always be achievable. If this is the case, engineers can take some steps to ensure that the accelerometer is fully in contact with the measurement surface. For curved cylindrical surfaces, some engineers will file a flat surface so the accelerometer will have full contact.

Others will use “modal guides” as seen in Figure 20.
 
These are small plastic pieces that are specially cut so that the accelerometer will maintain a consistent global coordinate system. These pieces can change the frequency content of the data measured, so it should be considered how these will affect measurements if an engineer chooses to use them. 

When mounted, it is also important that the accelerometer housing be electrically isolated from ground (Figure 21).
 
If an accelerometer is not isolated from the test object, ground loops can cause 50 Hz (Europe) or 60 Hz (United States) spikes to appear in the vibration data which originates from the electrical system, not from actual vibration.

More information:

• Ground Loops: What is causing that electrical noise?
• Modal Testing: A Practical Guide

4.2 Size: Accelerometer versus Test Object

If an accelerometer is mounted to a structure, the size of the accelerometer and structure need to be taken into consideration. For example, if a large accelerometer were mounted onto a small, thin piece of sheet metal, the accelerometer can add local mass and stiffening to the structure, which will change the modal properties and resonances of the structure during testing. 

In some applications, like modal testing, there are tools that can be used to account for mass induced frequency shifting.  More information in the knowledge article: Simcenter Testlab: Multi-Run Modal.

4.3 Cabling Considerations

The cable is typically the weakest point in a measurement system.  Cables are more likely to go bad or break than the accelerometer itself.

Some general cabling tips:

• Slack: Ensure that has enough slack and cannot be rattled or tripped on. 
• Tied Down: Make sure the cables are not able to whip around or get caught in any machinery during the test.
• Electrical: Cables should not be routed near power lines, otherwise they may pick up electrical line noise due to electromagnetic interference.
• Humidity: If working in a high humidity environment, create a drip loop that allows condensation to form and travel down the cable without allowing it to reach the sensor (see Figure 22).

Cable length can also have a direct impact on the accelerometer readings. Cables are designed to have small amounts of resistance so that a signal can travel through the cable unattenuated. The laws of resistance state that a conductive material’s resistance is directly proportional to the length, so as a cable increases in length, the cable’s resistance increases as well. Special graphs called nomographs can be used to select a cable length that will allow a user to measure specific frequencies without experiencing degradation, as shown in Figure 23.
 
Cables have a built-in capacitance that is also dependent on the length of the wire. The combination of capacitance and resistance in the cable forms an RC-circuit that acts like a low-pass filter. This means that as cable length increases, the amount of high frequency attenuation also increases, and data is lost. This is usually a problem at very high frequencies for very long cables.

More information on cabling:

• Long Cable Lengths and ICP Transducers
• Simcenter Testlab SCADAS and Long Strain Cables
• Simcenter SCADAS Hardware Cable Guide

5. Getting Started in Simcenter Testlab

Simcenter Testlab software with Simcenter SCADAS hardware can be used to measure vibration from accelerometers.  This section explains the basics of setting up a vibration measurement and calibrating the accelerometer in Simcenter Testlab.

5.1 Calibration


Direct YouTube link: https://youtu.be/IYlZIxenn1k


Before taking a vibration measurement, it is good practice to verify that the accelerometer is setup properly and working by providing a known calibration signal.  This is done even if the sensitivity value of the accelerometer is known and entered into the software.

Field calibration is done by mounting the accelerometer on a calibrator as shown in Figure 24:
 
5.2 Setup in Simcenter Testlab

Begin by opening the Simcenter Testlab acquisition package of choice (ex. Signature Acquisition, Impact Testing, Spectral Testing). Once open, navigate to the “Channel Setup” tab at the bottom of the screen as shown in Figure 25:
 

• In the second column of the channel setup, an “ON/OFF” checkbox is located next to each channel to determine if the channel will be on or off during measurement.
• Ensure that the channel group is set to “Vibration” and then name the channel in the column called “Point ID”. An example in figure 23 below shows the point ID for each channel called “Accel X” “Accel Y” and “Accel Z”. In this example, a triaxial accelerometer is connected with each direction being measured. 
•  Next, list the measurement direction in the tab labeled “Direction”. In this example, the direction of the channel for “Accel X” is “+X”. 
• The most important step is to select the conditioning of the channel. This step cannot be skipped because if ICP accelerometers are being used, they will not be able to measure any data. Under the column labeled “Input mode”, select the correct type of input for the accelerometer. If the accelerometer is an ICP, IEPE, or CCP accelerometer, choose “ICP” as the input mode. If a separate signal conditioner being used, select “Voltage AC.” 
• Ensure that the measured quantity is set to “Acceleration” in the “Measurement Quantity” column. 
• If the sensitivity value is known, enter the sensitivity values. If using a single axis accelerometer, enter the sensitivity value for the single direction. If using a triaxial accelerometer, enter the sensitivity value for each individual direction as they are usually all a little bit different. These can be found on the data sheets. 

For more information on using Simcenter Testlab for vibration measurements:

• Getting Started with Simcenter Testlab
• Simcenter Testlab Data Acquisition Tips
• Cool Channel Setup Tricks for Triaxial Accelerometers
• Simcenter Testlab Impact Testing
• Simcenter Testlab Signature

5.3 Setup in Simcenter Testlab Neo

To setup in Simcenter Testlab Neo, the process is very similar to setting up in Simcenter Testlab classic. First begin by opening the acquisition package of choice for Simcenter Testlab Neo such as “Time Data Acquisition”. Go to the Channels (tab at bottom left) as shown in Figure 27:
 

• In the "Home" ribbon at the top, there are several pre-defined channel views.  Use "Overview" to setup an accelerometer.
• “Input Mode” is called to “Conditioning” in Simcenter Testlab Neo.
• If TEDS are available, select the “Read TEDS” function in the lower left corner of Simcenter Testlab Neo. In Simcenter Testlab classic, it is a pulldown in the upper right.

• Simcenter Testlab Neo Introduction
• Simcenter Testlab Neo Time Data Acquisition