Two weeks ago, I was gluing an analog 3 axis accelerometer onto a heavy variable- speed motor and applying an offset weight to observe the vibration amplitude. I had run into a method to minimize vibration due to dynamic imbalance by measuring vibration amplitude: first without, then with an unbalancing mass, then with that mass moved 120 degrees forward, then 120degrees back from the reference - initial position. This is adequate to allow a simple code to determine where a mass should be placed and how big, to minimize dynamic imbalance. That was cute!
A week ago I was gluing a digital 3 axis accelerometer on a ceramic heater, to sample a few thousand values of its vibration both without, then with a tape partly obscuring its inlet, along with a few judicious taps here & there, in order to select a promising neural net provided by a vendor, in order to train an application to distinguish between "running well" and "beginning to fail". The more mature result would be to transmit a warning for the failing condition via bluetooth, or wifi. This machine learning application was to serve as my entry into the bottom end of AI: fairly useless in this case - but a shop with numerous machines might well appreciate advance warning of an impending failure and attendant $$$ out the window. That was cuter!
This week I stumbled into an application with educational possibilities..... in the pages of Wired..... when researching the answer to this Quora question, " How long would it take for a 4084 gram, 100 mm diameter steel ball to fall to the ground from a 100km altitude, taking in account altitude-dependent air pressure resistance, assuming 760 mm Hg and 15°C at sea level?
I was familiar with the iterated approach in Fortran from using it for forty years, but I would have stopped there.This time, I answered "About 246 seconds" after massaging a live graph which led to a short simulation which was modified and run and graphed by an interpreted language which looks like like VPython - a language with which I am entirely innocent - right there on my browser..
<https://imgur.com/a/rky9qTF>
This was too much fun to be a guaranteed copper-bottom correct response generated in just a few minutes.It reminded me strongly of that epic effort by John G. Kemeny and Thomas E. Kurtz and released at Dartmouth College in 1964 - the BASIC interpreter....which was lambasted in later years for its lack of rigor blah blah blah.Digital Equipment's Basic could be transported on a 4kB paper tape as I recall. I naturally bought a Commodore embodiment at Toys-R-Us in due course. In those days, when a HOME computer was being envisioned by IBM Atlanta (Just imagine a company with an 80% share of the commercial computer market casting around for suggestions!) I suggested a built-in BASIC compiler. Co-incidentally, they did just that!
Getting back on topic a small outfit developed TRINKET for educational purposes like this. I commend it to you.