MAGICO - WHY WE DO THE THINGS WE DO PART 2


“There are two different types of people in the world, those who want to know, and those who want to believe.” - Friedrich Wilhelm Nietzsche. This is the second instalment of the "Why We Do the Things We Do" in-depth series of articles by Magico. Alon Wolf writes: "How do physics shape the challenges that every speaker designer must face? What are the basic principles at play? And how do they drive engineering choices? At Magico, we are creating a series of articles to answer these questions."

Published in the first article the first topic, driver configuration identified the inevitable tradeoffs entailed in two approaches: D’Appolito/MTM driver arrangement and so-called driver time alignment. 

(Academic research[1] shows that the acoustic center of a source varies with frequency and distance from the source.) Our second subject, driver design, is far less controversial.

TOPIC 2: Driver design


We will explore drivers from the standpoint of four issues that every speaker designer needs to deal with – or ignore at their peril:
  • Drive linearity
  • Voice coil heating
  • Eddy currents
  • Diaphragm rigidity
This article will describe Magico’s comprehensive response to every one of these challenges. Readers looking for some “silver bullet” will be disappointed. Neither a single ingenious part nor a single conquered distortion defines a Magico driver. Of course, we were the first in the world to use Graphene in our bass and midrange cones (and more recently the first to use an aluminum honeycomb core in a sandwich between Graphene/carbon fiber skins). We were also first to apply chemical vapor deposition diamond to the surface of beryllium tweeter domes (and the first and only to produce a 28 mm Be dome). But these choices were the byproducts of a more comprehensive program: our no-holds-barred assault on the limits in every aspect of loudspeaker design.

[1] Jacobsen, F., Barrera Figueroa, S., & Rasmussen, K. (2004). A note on the concept of acoustic center. Journal of the Acoustical Society of America, 115, 1468-1473. https://doi.org/10.1121/1.1652036

The sound of Magico loudspeakers begins with the technology of Magico drive units.

Drive linearity


At the heart of a dynamic driver is the motor system, which translates the amplifier’s electrical signal into motion. The driver incorporates a coil of wire, the voice coil, suspended in a magnetic field that is concentrated in the air gap between a central pole piece and a donut-shaped top plate. As the audio waveform changes, current from the amplifier flows back and forth through the voice coil. The magnetic field pushes the voice coil back and forth. Connected to the voice coil, the driver’s cone or dome moves back and forth against the air, creating sound.

Voice coils generally use one of two configurations: overhung and underhung. Overhung means the voice coil is longer than the air gap. Underhung means the voice coil is shorter.

These drawings, not to scale, show the magnetic circuit in cross section. Overhung voice coils are longer than the air gap they move through. Underhung voice coils are shorter.

Loudspeaker engineers know[2] that the overhung voice coil has practical advantages including high efficiency, high maximum excursion, relaxed tolerances in the voice coil gap and dramatically lower cost. Thanks to low cost, overhung drivers are by far the most popular choice. Unfortunately, the force that the permanent magnet applies to the voice coil can drop as much as 50% as the voice coil moves away from the central, rest position. This drop-off can be asymmetrical.[3] Overhung voice coils are particularly prone to these problems, which sacrifice linearity. Motion does not faithfully track the input signal, generating substantial harmonic and intermodulation distortion. In addition, the overhung system distributes overall mass poorly, leading to unwanted pendulum effects and rocking modes. Inductance is also higher and more variable. These shortcomings are audible. Subtle details and microdynamics are lost. The speaker places a veil between you and the music.

For all these reasons, most Magico drivers feature an underhung voice coil. Maintaining constant magnetic force across the full range of voice coil motion, our underhung design comes much closer to perfect linearity – and substantially reduces distortion.

Our voice coil is also less prone to rocking. And the underhung design reduces inductance. All these differences deliver audible improvements in musicality, transparency and detail.

Having chosen underhung voice coils for their sound quality, we took extraordinary steps to elicit the best overall performance.
  • N52 Neodymium ring magnets. Typical drivers employ a single Ferrite or multiple Neodymium magnets, sometimes in “flower” configurations. We choose Neodymium because it delivers maximum energy (kilojoules) in minimum size (cubic millimeters). Where Ferrite has about 30 kJ/mm3, Neodymium achieves about 500 kJ/mm3. All our M and most of the S and A Series drivers use a single, super-sized N52 Neodymium ring magnet. Far more expensive than conventional magnets, the Neodymium ring generates much higher power and supports longer air gaps. The extreme compactness of Neodymium also enables better airflow. As an added benefit, Neodymium has superb coercivity, strongly resisting demagnetization in the presence of opposing magnetic fields all the way from room temperature to 446° F (230° C).
  • Longer air gaps. At a substantial expense, we increased the thickness of the top plate, to enable longer maximum excursion without sacrificing linearity.
  • Narrower air gaps. To concentrate the force of our ring magnet, we narrowed the gap between the pole piece and the top plate. This requires super precision in manufacturing and assembly, with unusually tight tolerances.
[2] Crocker, Malcolm J.; “Handbook of Acoustics,” John Wiley & Sons, New York, 1999, pages 1394-1395,
[3] Merit, Benoit; Lemarquand, Valérie; Lemarquand, Guy; Dobrucki, Andrzej; “Motor Nonlinearies in Electrodynamic Loudspeakers: Modelling and Measurement;” Archives of Acoustics 34, 4, 579-590, 2009
Our underhung motor system incorporates an extraordinarily powerful N52 Neodymium ring magnet. Thanks to these design choices, Magico drivers deliver the underhung advantages with just one tradeoff: substantially higher cost. 

Voice coil heating and how to overcome it


One of the most insidious, yet significant loudspeaker distortions occurs after a few seconds of music reproduction. Voice coils begin to get very hot, very quickly: as much as 106° F (41° C) in just one second, according to one classic research study.[4] As the voice coil heats up, its DC resistance can increase by 50% or even 100%. The standard frequency response sweep test occurs much too quickly to trigger this voice coil heating. So the effect passes undetected by classic lab measurements. But voice coil heating has a profound effect on loudspeaker sound. For one thing, the bass response changes, with reduced low-frequency extension and a midbass dip not present in the typical frequency response test. Just as important, the dynamic musical peaks are compressed as much as 3 dB.

Where the typical bass voice coil is 1, 1.5 or 2 inches in diameter, Magico voice coils are 5 inches (127 mm), providing much greater surface area for faster heat dissipation and better thermal stability. That makes a huge difference, contributing to highly accurate bass under all listening conditions. Magico loudspeakers deliver unconstrained musical peaks without the compression you will hear in conventional drivers.

[4] Button, Douglas J.; “A Loudspeaker Motor structure for Very High Power Handling and High Linear Excursion,” J. Audio Eng. Soc., Vol 36, No. 10, 1988 October, pages 788-796

The same approach applies to our midrange drivers. Where the typical midrange voice coil is 1 or 1.5 inches, Magico voice coils are 3 inches (75 mm).

A disassembled Magico bass driver showing the 5-inch (127-mm) voice coil and pure Titanium voice coil former.

We also looked beyond the typical materials for the “former,” the cylinder around which we wrap the voice coil. We sought out the optimum combination of light weight, high strength and appropriate electrical resistance (to reduce eddy currents). Conventional voice coil formers consist of paper, fiberglass or more recently Kapton® polyimide film. Kapton is extremely light and offers high electrical resistance, but is not strong enough. Considering stronger alternatives, aluminum is light and strong, but conducts electricity so well it’s used in wiring. We found our happy medium in pure titanium. While very expensive, titanium is so strong and light that it’s an essential component in jet engines. Yet unlike aluminum, pure titanium is moderately resistant to electricity, to help reduce eddy currents. We’ll discuss these eddy currents next.

Eddy currents and how to fight them


Magnetically induced eddy currents can be so powerful that trains use them as brakes. Just as eddy currents can stop a train, they can impede voice coil movement, degrading the sound. In dynamic drivers, the changing current from your amplifier creates temporary magnetic fields around the voice coil. These fields cause electrons to move in localized swirls, the eddy currents, in nearby metal parts. Eddy currents generate spurious magnetic fields that oppose the fixed magnetic field, slowing voice coil movement and creating distortion.

Because eddy currents are triggered by the interaction between the voice coil’s changing magnetic field and the surrounding metal parts, Magico has redesigned the surrounding parts to prevent eddy currents from forming. We use two techniques: maximizing magnetic saturation and minimizing inductance.

When the top plate and pole piece around the voice coil achieve total magnetic saturation, there’s no possibility of induced magnetic force, and hence no eddy currents. Conventional bass units drive the most critical areas of the top plate and pole piece to about 60% magnetic saturation. In the Magico woofer, we achieve 90-100% saturation.

In these cutaway diagrams of bass motor structures, dark red indicates full magnetic saturation. The Magico design (right) achieves visibly superior saturation.

Typical drivers, if they do anything at all to limit eddy currents, employ a copper “shorting ring” near the front of the magnet. At Magico, we encapsulate the entire voice coil gap with a precision-machined full copper sleeve. Our 127 mm copper sleeve reduces inductance by more than 90% (from 3.0 mH at 1 kHz to less than 0.3 mH). Magico’s magnetic circuit saturation, full copper sleeve and titanium voice coil formers all work together to reduce inductance in our woofers to just 0.085 mH – a remarkable feat. In this way, Magico drivers defeat eddy currents, suppressing another source of distortion and removing another veil between you and the music.

We encapsulate the entire 127 mm voice coil gap in a 0.15" thick copper sleeve. It decreases inductance by 90%.

Diaphram rigidity: the crucial importance of pistonic motion

At Magico, we believe that a loudspeaker cone or dome must move in unison like the rigid surface of a car’s piston. Achieving this pistonic motion means eliminating bending and flexing throughout the driver’s range of operating frequencies, the pass band. In addition, any breakup modes should either be damped or pushed well above the pass band – or both.

In this comparison of midrange drivers at 1,500 Hz, the red indicates inward motion and blue indicates outward motion. The paper midrange (left) shows a chaos of bending and out-of-phase motion where the Magico Nano-Tec midrange is uniformly pistonic. These images were generated with Klippel Laser Interferometry, which measures the motion of thousands of points across the driver surface. We selected 1,500 Hz because it’s right in the wheelhouse of midrange operation and squarely in the range where human hearing is most sensitive.

Non-pistonic motion generates out-of-phase sound and distortion unrelated to the music. Just as bad, the acoustic wave front emanating from a non-pistonic driver becomes incoherent, with out-of-control radial and circular modes. While these distortions will not show up on classic frequency response sweeps, they are massive in scale, clearly audible and simply unacceptable to us. And once the driver introduces these anomalies, the speaker is putting an inescapable and undesirable “fingerprint” on the musical timbre.

Like many ideals in audio, pistonic motion is simple to define, yet exceedingly difficult to achieve. At Magico, we pursue the ultimate in pistonic motion, with materials and configurations that establish new benchmarks in performance.

Nano-tec bass and mid-range cones


To achieve pure pistonic motion, our Nano-Tec® bass and midrange cones were the world’s first to use Graphene carbon nanotubes. Graphene takes the form of hexagonal lattices of carbon just one atom thick. Graphene combines incredible stiffness (elastic modulus) with the highest tensile strength of any material known to science. Where high carbon steel has tensile strength of 1.6 gigapascals (GPa), Graphene achieves 63 GPa, equivalent to 9.1 million pounds per square inch.

For even greater strength, most of our bass and midrange drivers eliminate the weakest part: the central dust cap. While the dust cap compromises pistonic motion, it greatly facilitates driver manufacturing. During assembly, the hole where the dust cap eventually goes provides easy access to the voice coil and pole piece. The critical centering of the cone/voice coil assembly within the magnetic gap is easy to achieve. Eliminating the dust cap means there’s no hole, no easy centering. Assembly time goes up and manufacturing yields go down. Nevertheless, we make this effort because the resulting cone is super-strong to reduce bending and flexing to the absolute minimum. In fact, our Nano-Tec cone diaphragm is so strong that, inverted on the floor, it will not deform under the weight of a person standing on it.

To keep break up modes well out of the pass band and attenuate ringing, our cones are self-damping. The Graphene carbon nanotubes form a sandwich around a super lightweight core. Many of our drivers use a foam core, while the latest are the world’s first with an aluminum honeycomb core between carbon fiber/Graphene skins. You get the closest possible approach to ideal pistonic motion.

Beryllium and Beryllium/diamond tweeter domes


Like paper cones, soft domes are constantly bending and flexing, distorting the music and disrupting the acoustic waveform. In our search for the ideal combination of light weight and high rigidity, we evaluated extremely stiff materials including aluminum, titanium and beryllium. Even in comparison to other advanced materials, beryllium is unmatched, with outstanding resistance to bending and flexing. These results are not surprising. For years, engineers have chosen beryllium for aircraft, spacecraft and Formula 1 engines (until the FIA banned it for conferring an unfair advantage).
Klippel Laser Interferometry analysis of total dome vibration at 10,000 Hz. Ideal performance would show an all-white dome in plan view and a completely smooth curve in cross section. While all three materials outperform silk domes, beryllium obviously comes closest to pure pistonic motion.

One material more pistonic than beryllium is diamond. Of course, diamond alone incurs a great weight penalty. This leads to unresolvable issues including low efficiency and severe rocking[5]. Magico was able to get some of the benefit of diamond, increased stiffness, with hardly any added weight. We were the world’s first to apply a carefully controlled layer of diamond by chemical vapor deposition on a beryllium dome. The combination of beryllium and diamond enabled us to go bigger, with the world’s first 28 mm beryllium dome. On the Magico M and S Series, diamond imparts super hardness to achieve pistonic motion of the highest degree.

1-inch diamond coated Beryllium dome tweeter of the Magico S7.

[5] Cardenas, William; Klippel, Wolfgang; "Loudspeaker Rocking Modes (Part I: Modeling)," AES Convention 139 (October 2015), Paper 9410

The price of performance
  • Magico’s unrelenting insistence on performance without compromise has led us to look beyond conventional driver solutions. And every refinement comes at a price.
  • Our underhung voice coil design uses tighter tolerances, more expensive parts and an extremely costly N52 Neodymium ring magnet.
  • Our pure titanium voice coil former is much more expensive than Kapton film.
  • Due to bigger metal parts and the tighter manufacturing tolerances required, our 127 mm (5-inch) titanium voice coil bass driver is far more expensive than the same driver with a 2-inch voice coil.
  • In the absence of the conventional dust cap, centering the cone and voice coil becomes substantially more time consuming and complex.
  • Outfitting our full copper sleeve costs several times as much as a copper ring and obviously much more than a driver that has no copper at all.
  • Our Nano-Tec cones are dozens of times the price of paper cones.
  • Our Beryllium dome costs much more than a silk dome, while our Beryllium/diamond dome is orders of magnitude more expensive.

The bottom line


We’ve seen the distortions caused by drive nonlinearity, voice coil heating, eddy currents and diaphragm flexing. Magico’s no-holds-barred assault on the limits of loudspeaker design has compelled us to overcome every one of these issues, regardless of cost. We use the latest in materials science; laser interferometry testing; extensive computer simulation and prototype testing of acoustic, mechanical, electromagnetic and thermal behaviors; plus meticulous assembly techniques and many more practices. In this way, Magico loudspeakers achieve musicality, transparency and microdynamics of the highest order.

Exploded view of Magico’s new cone, showing the aluminum honeycomb core.