Motor vibration measurement

Other Hardware Science and Measurement
1

Hi,
Im currently working on a project which requires testing the vibration of a rotating mass and calculating it's vibration velocity.

The assembly is turning at 3000 rpm and I need to extract the vibration in IPS.
I was intending to use the gy 521 sensor and a 328p arduino but I'm a bit worried about sample rate..
As I'm testing only one acceleration axis, a quarter turn at that speed is 5ms.
In order to calculate IPS, I need to measure the acceleration and multiple by time, adding the result to the previous one.
I was thinking of 10 times the "wave" length for oversample resolution which brings me to 500us per sample..
what do you guys think?

2

IPS - inches per second - may be industry standard, but why mix metric and imperial units?
However I understand Watts per metre squred isnt very intuitive.

3000 rpm = 50 rps = 20 msec per rotation.

If you are interested in measuring the peak amplitude you could use the nyquist criterion; So sample at intervals less than 10msec.
If you are interested in higher frequency components than that would guide your sampling rate.
It also depends on the resolution of the measuring instrument.

3

Hello liron_al

Take a text book read and study FFT for your project.

4

20ms is for a complete turn. If mass is out of balance, it will show twice as positive and negative G.
So actually I'm only interested in a single appearance of the wave, which means 10ms, not 5 as originally stated.
So all sampling calculations and storing the data needs to happen in 1ms..

5

Hi, @liron_al

A single appearance of the complete wave positive and negative, the complete vibrational movement, one rotation of the out of balance shaft, will occur once every 20ms.

Have you Googled;

how to measure vibration with arduino

Arduino Vibration Logger | Arik Yavilevich's blog.

Tom.. :grinning: :+1: :coffee: :australia:

6

True, but all I care about is 1/2 the turn as the next half should be identical but in negative values.
None the less, the sample rate must be 1ms.

As for the link for the vibration logger, I did read it but it is of no help for my case.
He does not state the rpm of the pump, also the term of "vibration level" is not a standard term..
I am looking for ips calculations.

The issue boils down to whether the 328p can handle sampling, float calculation and storing within 1ms

7

Then you will not be able to measure the vibration, if vibration at frequencies higher than 500 Hz is present.

Look up "Sampling Theorem" and "Nyquist Limit" to learn why.

8

Is the frame that is holding your rotating mass constrained so there is only a single plane of movement?

9

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.