Recently, the speaker system in my garage stopped working.  I’ve had speakers with integrated amplifiers in my garage for some years, and I’ve grown accustomed to the quality audio experience they provide.  Few things in life please me more than listening to Terry Gross’ Fresh Air or Nina Totenberg’s legal affairs correspondence while I’m running a 5 HP table saw through a three-quarter-inch piece of oriented-strand-board.  But my powered speakers had stopped working.  I needed a replacement, and soon.

Design Requirements

There are many speaker systems out there; it’s hard to know where to start one’s research.  I’d grown accustomed to a high-quality Bose Sound System I dug out of the dumpster so many years ago.  But the chances of me finding another discarded system in the trash-can of a rich guy were slim

Should I use a class-D amplifier?  Perhaps a class B? How should I decide?   What electromechanical drivers will I use?  I’ve never had any formal training in acoustic amplification and it is all completely overwhelming.

So like much of engineering — let me take a step back and start with the two critical design parameters.

• It needs to be future proof (e.g., last at least ten years without upgrade)
• It needs to provide two channels (Left and Right) so that I can listen to NPR’s “This American Life” the way Ira Glass meant it to be heard. A stretch goal will be to make it compatible with Jad Abumrad’s “Radiolab” and perhaps Peter Sagals’ “Wait, Wait, Don’t tell Me.”

Future-Proofing

Future-proofing is easy.  Most speaker dials go up to 10 on a 10-point-scale or 100 on a 100-point-scale.  But every currently designed system suffers from the same fatal flaw:  once the knob reaches the end of the scale, the volume stops increasing.

As long ago as 1984, Spinal Tap’s Marshall amplifiers were cleverly designed to go up to 11.  Knowing this, a prudent engineer might design a system that can go up to 12.  But I don’t want to redesign this project a year from now, so I’ll scale up to 13 and leave a bit of space to pencil a “.5” so the maximum range will be 13.5 on a 10 point scale.  That bit of genius is what separates tinkerers from engineers.

For safety’s sake, I will design a mechanical stop on the dial between 10 and 11 that consists of a small screw or dowel-pin.  Only under the close and careful scrutiny of adult supervision, with trained paramedics at the ready, will I risk removing the mechanical stop and enter the dangerous “lethal” zone that lies in the great unknown beyond the number ten.

Speaker Selection

Driver selection is a bit trickier, but fortune finds a way.

A few years ago, I attended the Consumer Electronics Show as a reporter, and I was costumed appropriately.  My outfit consists of an ill-fitting tan multi-pocket trench coat, an audio recorder, a clipboard, a camera, and a press-pass I had printed up and stuck in the hatband of my Fedora.

I find it essential to be readily identifiable in any profession, but people often pull me aside and tell me things when I dress like a caricature of a reporter, and CES 2018 was no different.

Most people associate CES with a wasteland of booths because that is what they see on television, and that is what most people experience.  But CES also provides private hotel suites where one-on-one interactions can occur.

I remember being on an upper floor of the Venetian, waiting for an appointment with AMS, when an engineer opened a door across the hall.  She sized me up and down, stopped, smiled, and said something I’ll never forget:

“Are you a reporter?”

Journalistic ethics bound me to discard my lit Cigar right there in the hallway, doff my hat, and respond in the affirmative and say “Yes ma’am I am.”

It’s kind of like when an undercover cop tries to buy grey-market integrated circuits from your private stash.  If you ask them if they’re cops, they have to tell you.

“Want to see some new speakers?”

“Speakers, eh?  Shouldn’t you have said ‘Want to hear something?” Engineers, above all others, know that fellow humans love to be corrected about trivial details.

There, along with three other people, she showed me Tectonic’s product line.  While these devices looked a bit like speakers, they were not loudspeakers, there was no cone.  No, these were something else called a Balanced Mode Radiators (BMR).  And I don’t need to tell you just how loud these things were — I’m sure you can tell from the picture below.

Image of one of Tectonics Balanced Mode Radiators from tectonicaudiolabs.com

What is the problem with loudspeakers?

Traditional loudspeakers have two fatal flaws.  I feel silly writing about it since you probably already know about breakup and directivity.  The general idea with any loudspeaker is to use as little electrical energy to move as much air as possible and have the response be consistent over a given frequency range and all directions in front of the speaker.

In speakers, a lightweight coil of magnet-wire is suspended in a permanent magnet.  As electrical currents pulse in one direction and then reverse, the energized coil is attracted to and repelled by the permanent magnet.  The coil is attached to a paper cone that pushes and pulls on the surrounding atmosphere.  The disturbances propagate outwards from the speaker cone into the environment where they reach our ears and are perceived as sound.

Most speakers consist of a copper coil suspended in the field lines of a permanent magnet. As current travels through the coil, the forces of magnetic attraction and repulsion are used to move an attached paper cone.

Breakup

An ideal speaker would be a perfectly rigid, massless driven piston.  Since such a thing doesn’t exist, thin paper is rolled into a cone to provide a low mass, semi-rigid piston.  The “semi” part of the rigidity is the problem.  The cones have vibrational modes that can establish standing waves at certain frequencies.  As you can imagine, if the cone’s overall displacement is positive while part of the cone’s local displacement is negative, or vice-versa, then the net displacement of that part of the cone will be zero, or perhaps even negative.

A finite-element-analysis for a fictional cone.    The result uses a color gradient to highlight areas of local displacement for one particular vibrational mode (selected based on a “coolness” factor compared to other results.)  Sound quality would degrade at frequencies much lower than the one shown — but this is the coolest looking vibrational mode I was able to identify.

A power-vs-frequency plot shows a marked decrease in speaker performance once vibrational-modes are established in the speaker cone.  Usually, engineers have to solve this problem by attenuating the signal at these higher frequencies and diverting the signal to a second smaller speaker called a “tweeter.”

Image of on-axis response taken from Tectonic BMRWhitePaper_Rev2.0_2019-1.pdf  Above ~150 Hz, the vibrational modes severely impact the speaker performance.

Directivity

At high frequencies, cones will focus the sound into a narrow beam.  Beam focusing is great for annoying neighbors but horrible for the astute audiophile who demands that Michael Barbaros “The Daily” sounds perfectly balanced no matter when in the garage they are working.

The Balanced Mode Radiator

As I noted above, cones are problematic.  They’re not strong enough to not resonate at audible frequencies, and they are not isotropic enough to uniformly bathe a room in the voices of award-winning journalists Alix Spiegel’s and Hanna Rosin’s “Invisibilia.”  So Tectonic did away with the cone.  BMRs use a coil to directly drive a large stiff plate.

Unfortunately, once you add the coil, the plates will have resonant frequencies that will affect performance.  The graph below shows pronounced dips at 2.5 kHz, 5.5 kHz, and ~18 kHz for a specific driven element.

This graph from Tectonics BMRWhitePaper_REV2.0_2019-1.pdf shows that a real element differs substantially from the theoretically ideal free disc.

The engineering comes when adding calculated ring masses that then largely remove the resonance modes.

The result is a speaker that has good performance over the entire range of human hearing.  There’s no need for crossovers, Woofers, Tweeters, etc…  There is just a single, carefully designed, pistonic element

Simulation of BMR in action from TectonicAudioLabs.com [[Alex — please use this image https://www.tectonicaudiolabs.com/wp-content/uploads/2018/11/bmranimation-crop.gif ]]

Summary

I demand nothing short of audio-perfection when listening to NPR over the drone of workshop machinery, and the BMR provides that perfection.  Click and Clack’s “Cartalk” reruns, here I come!