Taming The Wild Stirling Engine

24 Nov 2020

Introduction

Sensor technologies of all descriptions have intrigued me throughout my career. I have blogged about approximately 36 different sensor types on element14, but I have never used an industrial vibration sensor. I am really looking forward to trying out this Kemet vibration sensor and seeing what it can do.

To illustrate my interest in sensors, like vibration sensors, here is an example vibration sensing device I built for fun when I had a job working for a seismometer company. It looks like a miniature seismometer, but uses entirely different technology:

I was trying to make it very sensitive to ultra small but relatively quick motions of a desktop. This one is “de-tuned” a bit so it doesn't (usually) trigger when people walk by. It uses a simple brass disk with lead zirconate titanate (PZT) coated on one side. And a bit of a fancy circuit to stably boost sensitivity.

Kemet Vibration Sensor Experiments

For this experimentation challenge I will be performing some experiments to characterize the Kemet sensor so I can get a feel for its performance, but I also want to use it in a cool application. I will try to use interesting methods to keep the basic characterization tests entertaining and hopefully the main application I will explore is also interesting.

Sensor Frequency Response

I will be trying to characterize frequency response in a several ways:

a variable tuning fork type of vibration (the output can be compared to what a microphone picks up at the same time)
a sweep of vibration frequency - comparing excitation signal to sensor output at all frequencies in the time domain - and create a Bode plot
I will also try to inject both white noise and pink noise to see what the results look like in the time domain as well as the frequency domain (using an FFT)
I will also try injecting music into my drivers to see what the response will be to audio frequency vibrations, including trying to convert the samples from the sensor back into audio.

I have 2 different drivers I want to try, one is electromagnetic and the other is piezoelectric.

Some of these tests may highlight resonant frequencies in my test apparatus - we will see. (for sure the tuning fork experiment will showcase a resonant apparatus)

Resonant Frequency Exploration

I will try to excite the sensor's resonant frequency although the specs indicate this could be tough since it is way up in the ultrasonic range..

I will try with the piezoelectric driver even though I doubt it can stimulate that high a frequency, even with square wave harmonics
I will also try to stimulate resonance with an impulse impact

I don't expect too much from these experiments, since this is a strong suit of this sensor, but maybe I can confirm performance to some degree.

Sensor Sensitivity

I should easily be able to create small vibrations to subjectively explore sensitivity, however I am not sure yet how to quantify results. We will see what can be done.

The Wild Stirling Engine Problem

The application I want the vibration sensor to help with is a Stirling engine. Normally vibration isn't something you see, it is more like something you feel. The engine I have is highly unbalanced and it vibrates so violently that its whole chassis moves around on the table, even with rubber feet. This is a good visual and should make the videos less boring. I want to try different techniques to minimize vibration and stop it from wandering around. I will use the Kemet vibration sensor to quantify vibration with each mounting method and each balancing method to see which ones are practical and which ones do the best job of reducing unwanted motion. I don't just want to minimize vibration of the engine, I also want to minimize vibration of the table, so I will be using the vibration sensor on both the engine and the table.

Here is a quick video of my Stirling engine on the loose: