How to measure a fan's RPM without special equipment

[solarbird]

While we're at it - see previous post - here's a decent way to measure a fan's speed without having access to specialised hardware.

1. Extract the fan and provision it for power, but do not start it yet.
   
2. Ready a sound recording device of any type, such as your phone.
   
3. Using stiff but not overly stiff tape - I used packing tape - construct an inverted T type tab and adhese the top of the T to one side of the centre portion of the rotating fan blade assembly so that the extended tab spans the radius on one side of the fan, and the tab portion sticks out, away from the fan. If any part of your assembly has any possibility of hitting any other part of the fan, you're putting it in the wrong place. The tab you've made should rotate with the fan blades, once per revolution of the blade assembly.
   
4. Using the same tape, construct a lightweight rod of, oh, 5cm length. Exact length doesn't matter.
   
5. Start the fan.
   
6. Position the rod so that it is continually struck, lightly, by the previously-made tab adhered to the fan. It should make a ticking sound, like a playing card in bicycle wheel spokes. Each tick sound will represent one (1) rotation of the fan.
   
7. Start the sound recorder, and record some number of seconds, making sure you get as light an impact on the tab you constructed as possible while making sure the tick sound remains clear. (This is so you slow the fan down as little as possible or not at all.)
   
8. Save the recording, then load it into any audio software of your preference that will let you display the waveform and zoom into a one second section thereof. Audacity is one such free software package, but there are many others.
   
9. Looking at the waveform, you will see a series of vertical lines, each one of which is a tick noise, which should be louder than background/air noise in the recording. If you do not see (or hear) this, you have a bad recording.
   
10. Count the number of these lines contained within exactly one second of time. That gives you revolutions in that second. If you are comfortable that this one second is representative, multiply by 60 for revolutions per minute (RPM). Alternatively, count across several seconds and do math accordingly. It's up to you.

Comments

[mdlbear]

Do you think Audacity's FFT display would make this any easier?

[solarbird]

Maybe? If you can get a good consistent length of time sample then probably sure. I don't know how sensitive Audacity's FFT is to stuff like, hm... variable loudness and such, and also some variation in speed depending upon how consistent you can get the tick pressure on the tab, right? But it's probably gonna be kind of flexible with that since it's an audio app. So yeah, that should work, if you don't want to count ticks for some reason.

Give it a go, see what it does ^_^

[wrog]

Save the recording, then load it into any audio software of your preference that will let you display the waveform

Or you could, say, listen to it. 3 octaves above A-440 is 3520 s⁻¹ which is the A at the top of the piano keyboard, which is still quite audible. 3000 will be a G that's a quarter tone flat.

And if you've got a tuning fork (or maybe an actual piano that's been tuned vaguely recently), you don't even have to have absolute pitch

[solarbird]

But it's not a tone - at least, not to me. The spikes aren't sustained enough to make a proper square wave. They're very short - since this fan has seven blades (iirc) then it's no better than a 7-1 signal-spike-to-zero ratio, or however you'd say that.

For me, at least, it's a buzz where you clearly hear the individual hits and it doesn't resolve into a meaningful pitch. The individual impacts have a tone, of course, but that's different.