Measuring the Frequency and Amplitude of the vibration of a cantilever beam

I am currently working on a Free and Foreced Vibration apparatus. In the experiment there will be motor with unbalanced weight attached slightly offcenter to the cantilever beam and when the motor rotates the beam will oscillate and it will create vibration. My purpose is to measure the Amplitude(of the beam) and frequency of the vibration. Should i use any laser distance sensor or any accelerometer attached to the beam. If i use ADXL345 how can i measure the amplitude and vibration from the acceleration data's?
Free-and-Forced-Vibrations-Apparatus

Cant you use a strain guage?

I think you have to integrate acceleration over time.
For short test duration it should be ok, for longer ones I think the '0' may shift ( but as oscillation is always centered to '0' I think you can derive an offset correction algorithm )

Or you can use a strain gaige sensor ( similar to those used in load cells, afterall your is a load cell )

Laser sensor are quite 'slow' compared to the other two methods
I have a keyence IL-600 which max sampling frequency is 0.33mS ( but quite noisy without additional averaging )

The maximum sampling rate of the ADXL345 is 3200Hz. Is the fast enough for your set-up?

yes..my system will only vibrate between 50-400Hz...maybe i have to use FFT..but how can I do that onboard

If this is any type of commercial application I would be looking at a commercial accelerometer.

Vendors like B&K, Endevco or similar. You can measure frequency and amplitude. You should use a charge amplifier with these but some include an onboard amp. You get an image that looks like this.

Vibe Fan 1

Make sure your choice can respond to the frequency range mentioned. The pictured accelerometers use a 10-32 screw base or you can use a magnetic base. If you want accurate results correct mounting is important.

Just My Take
Ron

I can't help with the data analysis but if you use the ADXL345 and your data bandwidth is 400Hz, then you should use the SPI interface and a sampling rate of 1600Hz or 3200Hz. If you want to collect 1 second of data at 1600Hz you will need at least 1600 x 6 bytes = 9600 bytes of RAM.

The Uno only has 2048 and the Mega only 8192, so neither has enough RAM.
You can buy an Arduino with more RAM or maybe use external RAM like SPI/I2C FRAM

I'm a 'super fan' of esp32 ( so consider this ), but at a bargain price ( if it is important ) you have a 240MHz dual core processor with 500Kb ram 4-16Mb flash... a variety of boards available... fft should not be a problem ( did I mention wifi and bluetooth? ).
It could even be used as the main processor of the whole system

1 Like