Previous Page   Next Page

 

Frequency analysis in the industrial field using Freescale accelerometers



 


Mechanical systems such as electric motors tend to vibrate, even during normal operation. An unexpected change in the system’s vibration signature may indicate a potentially-damaging operating condition – such as an unbalanced load in a washing machine – or may be a precursor of failure.

MEMS accelerometers such as the Freescale MMA7361L can be used to collect vibration data for an embedded vibration-analysis algorithm supporting value-added features such as operator-safety warnings or failure prediction. This Design Note describes how accelerometer data can be manipulated to analyse the performance of various domestic and industrial products.

 


 

Types of Vibration

There are two types of vibrations: free vibration and forced vibration. Free vibration occurs when a mechanical system receives an initial stimulus and then vibrates freely; a tuning fork is a good example. The mechanical system will vibrate at one or more of its natural frequencies and eventually dampen down to zero.

Forced vibration is when an alternating force or motion is applied to a mechanical system; an example is the shaking of a washing machine due to an imbalance. The frequency of the vibration is the frequency of the force or motion applied. The order of magnitude is dependent on the actual mechanical system.

 


Fig. 1: Load condition: centred light weight.

 

Data acquisition and system analysis

The sensor used in this example is the Freescale MMA7361L analogue-output accelerometer. The data from the accelerometer is collected and data-logged in time.

The transducer output undergoes signal conditioning inside the accelerometer. The typical Freescale low-g consumer accelerometers have a bandwidth response between 300Hz to 400Hz at the -3dB point. The accelerometer can detect signals above 400Hz but the internal low-pass filter will limit response as the amplitude decays. About 10% is seen at 1kHz.

In this example the bandwidth is set to 1024. The sample rate is chosen to be double the Nyquist frequency, which is the highest frequency detectable.

The time-domain data can be useful for doing waveform analysis of impulse or transient vibration signals, but it can be difficult to determine vibration faults in a mechanical structure.

The frequency of a simple vibration can be calculated from its period. The vibration signals from most rotating machinery contain harmonics of the fundamental rotation frequencies, so data has to be analysed by Fourier methods, which are established for periodic and random vibrations using a Fast Fourier Transform (FFT) algorithm. It is therefore necessary to process the datalogged time domain data and convert it into the frequency domain for spectrum analysis.

 


Fig. 2: Heavy load out-of-balance condition.

 

Simple DC brushed motor example: spectrum analysis

In Figure 1, a small weight was centred on the shaft of the motor by hand. It is noticed that the dominant frequency changed from 50Hz to 66Hz from the no-load conditions, and few harmonic frequencies are visible. The addition of the weight caused the frequency of the motor to reduce.

In Figure 2, a weight has been placed on the shaft to create an out-of-balance scenario. Due to the asymmetry of the mass, there are centripetal accelerations which are causing the motor to move resulting in large-amplitude accelerations. The fundamental frequency has shifted significantly, to 25Hz. Harmonics are present on the 2nd, 3rd, 4th and 5th frequencies, and above.

 

Consumer and industrial vibration applications

There are many common devices that create vibrations and can be analysed using an accelerometer. The accelerometer can be useful to determine the difference between normal characteristic conditions versus potentially dangerous or problematic situations. Indeed, it is very helpful in certain vibration applications to be able to predict failure or hazardous vibration conditions.

The Freescale MMA7361L accelerometer has been used to analyse equipment ranging from musical instruments, household appliances, personal care products, power tools and a compressor system. Many of these analyses are included in application note AN3751, available at: www.freescale.com/files/sensors/doc/app_note/AN3751.pdf

 

  Freescale / MMA7361L

 

 

 

Previous Page
Terms of Use  |  Privacy Policy
© 2010 Future Electronics. All rights reserved.

Next Page