Brass And Nickel Work Together In This Magnetostrictive Earphone

When you go by a handle like [Simplifier], you’ve made a mission statement about your projects: that you’ll take complex processes and boil them down to their essence. So tackling the rebuilding of the humble speaker, a device he himself admits is “both simplified and optimized already,” would seem a bit off-topic. But as it turns out, the principle of magnetostriction can make the lowly speaker even simpler.

Most of us are familiar with the operation of a speaker. A powerful magnet sits at the center of a coil of wire, which is attached to a thin diaphragm. Current passing through the coil builds a magnetic field that moves the diaphragm, creating sound waves. Magnetostriction, on the other hand, is the phenomenon whereby ferromagnetic materials change shape in a magnetic field. To take advantage of this, [Simplifier] wound a coil of fine copper wire around a paper form, through which a nickel TIG electrode welding filler rod is passed. The nickel rod is anchored on one end and fixed to a thin brass disc on the other. Passing a current through the coil causes the rod to change length, vibrating the disc to make sound. Give it a listen in the video below; it sounds pretty good, and we love the old-time look of the turned oak handpiece and brass accouterments.

You may recall [Simplifier]’s recent attempt at a carbon rod microphone; while that worked well enough, it was unable to drive this earphone directly. If you need to understand a little more about magnetostriction, [Ben Krasnow] explained its use in anti-theft tags a couple of years back.

24 thoughts on “Brass And Nickel Work Together In This Magnetostrictive Earphone

    1. In the article there’s a picture of one of my early prototypes which uses a wooden diaphragm, and it works just as well. In fact, my earliest prototypes were simply a coil of wire wrapped around a steel rod, with the end of the rod pressed against a thin wood panel. You could probably make one out of a nail and a piece of cardboard, honestly.

      1. Same comment: if the wire is responding to the magnetic field by attraction in any way, you’ll have a heck of a time centering it so perfectly that it won’t vibrate the diaphragm from that alone.

          1. Or the steel rod from being physically moved around by the magnetic field.

            Remember, the point is that the rod should change shape, not location. Putting a steel rod in a magnetic solenoid with AC current will vibrate the steel anyhow.

  1. If the magnetic field causes the rod to shrink, for example, why doesn’t the pitch of the sound double due to the magnetic field going to a maximum twice for each cycle (n-s) then (s-n)? Think waveform of full wave rectified AC before being filtered.

    1. That honestly confused me as well. This type of speaker sounds fine on an AC signal when it really shouldn’t at all. Maybe there’s some sort of hysteresis effect where the rod stays biased in one direction? Or maybe that particular type of distortion is hard to detect by ear? It’s puzzling.

      1. A simple suggestion (face palm … sigh) take the sound (original) run it through an FFT then take the recorded sound from the transducer and run it through the FFT you can adjust volume and take the spectral difference to see what is or isn’t happening that way.
        Don’t worry using Fourier series simplifies things… really!
        You can use a class AB amplifier with your transducer and use a standard summing ampto to bias the field in either direction and read the spectra output as you adjust the audio output. Just be sure to have a +-6V supply +-2.5V supply etc. to power it and adjust the bias. Most times audiophiles are removing offset voltages in this case it may be useful to see what happens as you add offset or subtract offset.

    2. Doubling the frequency goes up one octave, so it still sounds the same and the brain infers the missing fundamental.

      It’s essentially the same effect as filtering out the lower frequencies through a speaker that doesn’t produce good mid/bass frequencies, so for instance 300 Hz is missing but 600 Hz gets through – it sounds a bit tinny but otherwise perfectly normal.

      1. Furthermore, what’s happening with the magnetostrictive rod is that only the fundamental frequency and maybe the second harmonic is doubled. The higher harmonics that ride on top of the fundamental aren’t strong enough in amplitude to flip the polarity of the signal, so they get reproduced normally through each half-cycle of the fundamental frequency. It’s as if the fundamental frequency acts as a DC bias on the coil, and since the brass disc can’t reproduce low frequencies anyways the doubled fundamental gets filtered out.

        Fun fact: human vocal cords don’t actually vibrate faster than 200 Hz in most cases. Yet human speech goes up from 70 to 3000 Hz. How come? Because the vocal tract and mouth form overtones which are excited by the lower frequencies generated by the vocal cords, and these higher frequencies are heard as speech.

        1. On vocal chords: While I think the gist your comment was correct, the phrasing might lead to conclusions that are not.

          The vocal tract does not “form overtones.” The vocal chords produce a wave shape that is, for lack of a better description, a heavily-rounded saw tooth. It is therefore rich in harmonic content.

          Collectively the throat, mouth, and nasal passages function as mechanical band-pass filters and resonators. The vocal tract performs what is called “formant filterting.” Formant filtering both accentuates and attenuates portions of spectrum-rich content generated by the vocal chords…which results in voiced vowel sounds.

          The cool thing about our formant filters is that the points of resonance are continuously adjustable. We can slide them up and down on the fly. This allows us to produce complex vowel sounds like “I” which is actually “ah-ay-ee.”

          In the 1960’s, long before personal computers and speech-synthesis chips/software, Bell Labs produced an instructional kit to teach the basics of electronic speech synthesis. Example:

          https://beatriceco.com/bti/porticus/bell/belllabs_kits_ss.html

          Fun fact: LTSpice has a function that allows you to export a modeled circuit signal to a WAV file. At one point I modeled the circuitry in the I just described kit and was able to hear the resulting “A,” “EE,” “O,” “OO,” and similar vowel-like sounds with my own ears.

          I created my own version of Bell’s filters, replacing inductors and such with op-amp circuitry. A simple astable flip-flop (square waves are rich in odd-order harmonics) seems to make as good a set of vocal chords as anything else.

          This is fun stuff to play with.

          1. >The vocal tract does not “form overtones.”

            Yes it does. It literally forms the overtones – that’s why the filter is also called a “formant filter”. In non-linear systems, signal energy can also shift from lower to higher frequencies.

    1. Not quite. I used that shape because it’s convenient for this application, but old telephone handsets were pretty much universally electromagnetic, not magnetostrictive. There was a magnetostrictive speaker made by Philipp Reis in the 1860s, but it had a different shape and output mechanism (I’m currently experimenting with this). Like most of what I do, the earphone I made is sort of an alternate/unexplored branch of the historical tech tree.

  2. I wonder what the weight of this type of speaker is compared to a traditional electromagnetic type with similar characteristics? Maybe this could be a good way to make lightweight speakers?

Leave a Reply to ThopterCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.