Detecting machine anomalies at higher frequencies! 16kHz wideband MEMS accelerometer "ADXL383"

Narrow down by specifying conditions

現在2207件がヒットしています。check

ProductsPickup SensorsAnalog Devices

To prevent equipment failures, there is growing interest in "condition-based monitoring (CbM)," which involves constantly monitoring the condition of motors, pumps, fans, Box, and other components. Condition-based monitoring utilizes various information such as temperature, sound, and current, but "vibration" is particularly important for understanding the condition of rotating machinery.

Analog Devices also considers vibration sensing to be an important measurement tool for monitoring and detecting potential problems in rotating machinery.

The performance of condition-based maintenance solutions is determined by vibration sensors | Analog Devices

This time, we'll introduce the ADXL383, a 3-axis MEMS accelerometer that achieves a wide bandwidth of up to 16kHz.

This article will explain the differences between this model and the ADXL382 in the same series, as well as why a wideband accelerometer is necessary for fault prediction.

Why "broadband" is required for acceleration sensors used for fault prediction.

When a malfunction occurs in rotating machinery, vibrations different from those under normal conditions are generated.

For example, bearing damage, shaft misalignment, imbalance, and mechanical looseness each manifest as different vibration spectra. Therefore, by acquiring vibrations with an acceleration sensor and analyzing the frequency components, it is possible to understand the condition of the machine and any signs of abnormalities.

One of the sensor performance characteristics that becomes important here is "bandwidth."

Initial bearing failures manifest as high-frequency components.

When defects such as cracks or scratches occur in a bearing, an impact is generated each time the rolling element passes over that defective area.

In the initial stages of failure, this shock generates relatively small-amplitude, high-frequency vibration components. Analog Devices' technical documentation indicates that high-frequency components exceeding 5 kHz may appear as an early sign of bearing failure.

Choosing the Optimal Preventive Maintenance Sensor | Analog Devices

As the malfunction progresses, the vibrations on the low-frequency side also increase, making them easier to detect even with more common accelerometers. However, at that stage, a complete failure may already be imminent.

Furthermore, impulses generated by bearing defects contain high-frequency components, and initial signs of failure can range from a few kHz to over 10-20 kHz.

In other words, low-noise, wide-bandwidth accelerometers are crucial not only for detecting the malfunction itself, but also for capturing early signs of anomalies.

Wide bandwidth is also important for detecting gear anomalies.

Formula for calculating gear meshing frequency

Broadband vibration measurement is also important for monitoring the condition of gears. Even in normal gears, the meshing of the teeth generates vibrations called "gear mesh frequency." The gear mesh frequency is expressed by the following relationship:

Gear meshing frequency = Shaft rotation frequency × Number of gear teeth

Therefore, even if the shaft's rotational speed itself is not very high, the observed frequency can rise to several kHz depending on the number of teeth. Furthermore, if a localized failure occurs, such as a crack in the teeth, a short-duration impact will generate low-amplitude, broad-bandwidth vibrations.

Analog Devices also explains that a wide measurement bandwidth is crucial in detecting gear failures because the number of teeth acts as a multiplier in the frequency domain.

Thus, in order to monitor the condition of bearings, gears, and other components in more detail, it is important not only to measure the magnitude of vibrations but also to be able to acquire vibration information up to high frequencies.

ADXL383: A wideband 3-axis accelerometer with a maximum frequency of 16kHz.

As explained above, in order to monitor the condition of bearings, gears, and other components in more detail, it is important to be able to acquire vibration information up to high frequencies.

Therefore, the product I'd like to introduce today is Analog Devices' 3-axis MEMS accelerometer, the "ADXL383".

ADXL383 is ±15g, ±30g, ±60g Digital output corresponding to the measurement range 3 shaft MEMS This is an accelerometer. Its main feature is its high performance. HP) In mode maximum 16kHz broadband It supports this feature.

Furthermore, it achieves low power consumption of 520μA in high-performance mode and 33μA in ultra-low power consumption mode. In addition to SPI and I²C, it also supports audio data output interfaces such as I²S, TDM, and PDM.

Key application examples for the ADXL383 include condition-based maintenance, structural monitoring, seismic imaging, robotics, and audio/active noise cancellation (ANC).

Comparing ADXL382 and ADXL383 - What's changed?

To make the features of the ADXL383 easier to understand, we will compare its main specifications with those of the ADXL382, which belongs to the same ADXL38x family.

Item	ADXL382	ADXL383
Number of axes	3 axes	3 axes
Measurement range	±15g / ±30g / ±60g	±15g / ±30g / ±60g
Built-in ADC	16 bit	16 bit
Maximum bandwidth	8kHz	16kHz
Midband noise density XY	44μg/√Hz	44μg/√Hz
Midband noise density Z	55μg/√Hz	55μg/√Hz
High-performance mode current consumption	520μA	520μA
Ultra-low power consumption mode current consumption	33μA	33μA
SPI / I²C	Correspondence	Correspondence
I²S / TDM / PDM	Correspondence	Correspondence

The ADXL382 and ADXL383 share many basic specifications, including measurement range, built-in ADC resolution, current consumption, and various digital interfaces.

On the other hand, a major difference is the maximum bandwidth. While the ADXL382 has a maximum bandwidth of 8kHz, the ADXL383 extends it to a maximum of 16kHz.

As mentioned earlier, initial failures in bearings and gears may show signs of abnormality on the high-frequency side. Therefore, the ADXL383 can acquire vibration information in a higher frequency range than the ADXL382, expanding the frequency range that can be analyzed by condition monitoring.

It should be noted that simply increasing the bandwidth from 8kHz to 16kHz does not mean that "fault detection performance will double." Actual fault detection performance depends on various conditions such as the vibration characteristics of the target equipment, the sensor mounting method, noise, and the analysis method.

A major advantage of the ADXL383 is that it can now observe vibration information above 8kHz, which was previously outside the measurement bandwidth.

How does the ADXL383 achieve a wide bandwidth of 16kHz?

Another feature of the ADXL383 is that it achieves a wide frequency bandwidth by utilizing digital signal processing.

The frequency characteristics of the sensor are corrected using a digital EQ.

Because MEMS accelerometers have mechanical resonance characteristics, the frequency response of the sensor element itself is not perfectly flat.

The ADXL383 allows you to use a 4th-order digital equalizer (EQ) filter in high-performance (HP) mode.

This EQ filter corrects the sensor's frequency response, extending the measurement bandwidth up to 16kHz. The EQ filter is optimized for an output data rate (ODR) of 32kHz.

In other words, the ADXL383 achieves a frequency response that allows for use at higher frequencies by applying digital correction to the signal obtained from the MEMS sensor.

The ADXL383 datasheet states that when digital correction is used, the relative flatness in the frequency range below 15kHz is 1.6dB (XY axis) and 1.9dB (Z axis).

Frequency response of ADXL383 before applying digital EQ

Frequency response before applying digital EQ

Frequency response of ADXL383 after applying digital EQ

Frequency response after applying digital EQ

The trade-off between wider bandwidth and noise

On the other hand, it's not necessarily true that "the wider the bandwidth, the better."

The ADXL383 datasheet indicates that when the signal bandwidth is extended to 16kHz in HP mode, high-frequency noise components are also included in the signal bandwidth, resulting in an increase in noise density as you approach the upper end of the bandwidth. The noise density of the ADXL383 is 44μg/√Hz (XY axis) and 55μg/√Hz (Z axis) around 8kHz, but becomes 171μg/√Hz (XY axis) and 208μg/√Hz (Z axis) in the full 16kHz bandwidth.

Therefore, in actual applications, "16kHz Rather than "use until," It is important to consider the balance between the fault frequency you want to observe and the required noise performance. is.

For example, if the necessary vibration information is concentrated below 10kHz, you can consider using it to suppress noise by limiting the unnecessary high-frequency band. The ADXL383 also has built-in programmable LPF (low-pass filter) and HPF (high-pass filter).

Relationship between ADXL383 bandwidth and noise density

The ADXL383 is actually also suitable for audio applications.

Up to this point, we've mainly discussed condition-based maintenance, but the ADXL383 has another interesting feature.

ADXL383 is SPI or I²C but also I²S, TDM, PDM It supports audio data output interfaces such as those mentioned above.

I²S, TDM, and PDM are data interfaces widely used in digital audio equipment. Some may find it a little unusual to see these interfaces on an accelerometer. However, accelerometers measure the "mechanical vibrations" of objects. With sufficient bandwidth and low noise performance, accelerometers like the ADXL383 can capture not only vibrations from mechanical equipment but also mechanical vibrations in the audio frequency range as signals.

Furthermore, because data can be output in I²S, TDM, and PDM formats, it is easy to incorporate the acquired vibration data into the data paths of MCUs, DSPs, and SoCs that support audio signal processing, allowing the acquired vibration data to be handled by the same signal processing systems as audio signals.

Analog Devices describes the ADXL383 as being able to measure audio signals and heart sounds with high accuracy even in high-vibration environments, and cites audio and active noise cancellation (ANC) as examples of applications.

One of the key features of the ADXL383 is that, in addition to being a broadband acceleration sensor for condition-based maintenance, it can be deployed for various sensing applications that utilize vibration.

Application example

Taking advantage of the ADXL383's features such as its wide bandwidth of 16kHz, 3-axis measurement capabilities, and low power consumption, it can be applied to the following types of applications:

Condition-based maintenance and predictive maintenance

In rotating equipment such as motors, bearings, gears, pumps, and fans, changes in mechanical condition can be detected by monitoring the vibration spectrum.

In particular, the 16kHz wide bandwidth is advantageous for monitoring the condition of bearings, gears, and other components that require observation at higher frequencies.

Structural Monitoring Robotics

The ADXL383 can acquire vibration and acceleration data in three axes, making it suitable for applications such as vibration monitoring of equipment and structures, and vibration and shock sensing in robots.

Because the ADXL383 supports I²S, TDM, and PDM, it can be used for vibration sensing applications that are slightly different from typical accelerometers, such as audio signals, ANC, and heart sounds.

Audio and vibration sensing

Because it supports I²S, TDM, and PDM, it can be used for vibration sensing applications that are slightly different from typical accelerometers, such as audio signals, ANC, and heart sounds.

Considering a rapid deployment of ADXL383 using an evaluation board.

Furthermore, to facilitate the evaluation of the ADXL383, Analog Devices provides the "EVAL-ADXL383" evaluation board. The EVAL-ADXL383 allows for quick verification of the ADXL383 's features, such as its wide bandwidth of up to 16kHz and low noise, using the actual hardware. The lineup includes the "EVAL-ADXL383-1Z" for SPI interface evaluation and the "EVAL-ADXL383-2Z" for I²C interface evaluation.

EVAL-ADXL383-1Z for SPI

EVAL-ADXL383-2Z for I2C

Summary

This time, we introduced the ADXL383, a 3-axis MEMS accelerometer that supports a wide bandwidth of up to 16kHz.

In condition-based maintenance, it is important to detect signs of failure as early as possible, rather than waiting until the machine has completely broken down before detecting any abnormalities.

In cases of initial bearing failures or gear malfunctions, high-frequency components can provide crucial information, and acquiring this information requires a high bandwidth and noise performance from the accelerometer.

The ADXL383 shares the same measurement ranges of ±15g / ±30g / ±60g, low power consumption, and various digital functions as the ADXL382, while extending the maximum bandwidth from 8kHz to 16kHz. Furthermore, it supports frequency response correction with digital EQ and is compatible with I²S/TDM/ PDM in addition to SPI/I²C, making it suitable for a wide range of vibration sensing applications.

If you need to acquire vibration data at higher frequencies for machine condition monitoring or predictive maintenance, the ADXL383 is a strong option, so please consider it.