This is a real-time project where I used ADXL335, a 3-axis-accelerometer to measure unknown vibrations and then analyze them in Matlab and plot them in Excel
I uploaded the project video as this Linkedin post and also uploaded in the following YouTube Video You can check it for the demonstration and... Enjoy!!!
The main task of this project is to find the car's (test object) vibration and analyze all the outputs. With a similar setup and codes, any kind of mechanical vibration of an object can be found.
- For understanding about the mechanical vibration analysis, this paper is very much useful: Measuring-mechanical-vibrations.pdf
- To know how to work with ADXL335 sensor, this LastMinuteEngineers tutorial is very much helpful.
- Check the ADXL335 datasheet
- Arduino
- ADXL335 3-axis accelerometer (minimum 2 for comparing outputs)
- long wires (I used cat6 cable)
- veroboard, male headers, jumper cable, soldering iron, etc.
- 5V power supply, can be a power-bank, or battery. (I used the laptop's USB output for powering the setup)
- Vibration speaker (I didn't have any. but it is very useful for checking the results for a given frequency)
Circuit Diagram is made using circuito.io an easy readymade drag-and-drop circuit design app
The whole device looked like this. Here I have used CAT6 cables for covering long distances.
To obtain the task, first, the sensors' data is collected from Arduino IDE's serial monitor. Then it is stored as a CSV file and MATLAB will read that CSV as a table. Then by applying gaussian smooth and bandpass filter the noises from the signal were reduced and the signal now looked like this:
then applying fast fourier transform the signal will shift from Time Domain to Frequency Domain. The frequency that has the peak power will be the naturally measured frequency and the rest of the frequencies are the noises. You can check the imp_outputs folder's images for the details outputs.
Finally, another custom method for finding frequency is done by Excel in here. For different tire pressures, different frequencies were found for both front and back sensors. The output of the final figures for different pressures is
from all the figures and the graphs the vibration of the car (where the sensor is placed) was varied for several reasons including:
- acceleration and deacceleration of car
- noise of the sensors ( for the higher frequency we can see the noises )
- road defects and others