Genelec and Ribbon technology - Forum

There are many kinds of drivers, which are called ribbons without actually being so. The origin of a ribbon transducer is a microphone, where an extremely thin aluminum foil, 1…2 μm thick, a few mm wide and a few cm long, is suspended from its ends to a strong transversal magnetic field. As it is very light, it moves with sound and its movement in a magnetic field induces voltage, which is raised to usable levels with a transformer.

Ribbon loudspeaker is reciprocal of ribbon microphone, only larger. Due to dimensions of the air gap the ribbon loudspeakers are used as tweeters. As the ribbon impedance is very low, in the milliohm region, it needs a matching transformer. Inherently the ribbon transducer is expensive, as it needs strong magnet and the said transformer, which is not cheap either. The positive side is that the frequency response is very wide towards the upper end of the spectrum (up to 100 kHz, depending on the transformer) and all sorts of distortions can be very small. Hence the sound quality is very high. Also the sensitivity can be good, actually in the range of 92...97 dB/W. Ribbon materials are usually aluminum and sometimes beryllium. Response ripple can be extremely small.

As the mass of the ribbon is very small, a few milligrams only, its thermal capacity is low. In other words it heats easily and the heat is dissipated usually through convection to outside air and radiation. To increase the power handling, the ribbon should be larger and exposed to free air. Larger area means increased directivity, which may or may not be desirable. As is usual, some kind of compromise needs to be done, depending on the desired application. Long ribbons are definitely possible and manufactured, if line sources are needed. Midrange ribbons can be made but they are very rare and very expensive due to magnet structures. Also supporting a long ribbon for heavy use can be a problem.

The Genelec ribbon, as used in the S30 series and in the 1022A, is a real ribbon made with these principles. The benefits are, as you mentioned in your question, very high sound quality, very flat and wide response and negligible distortion. The limiting factor is power handling. It was perfect to the output category of S30 and 1022A but squeezing more output would degrade reliability. It is true that the impedance matching of the ribbon to air is good. This is due to the very low mass per unit area. The same is true with electrostatic drivers where the foil is very thin and light. However, the correlation to sound quality is not so direct. Any radiating surface following exactly the input signal does the job, special attention on the words "following exactly." Ribbon is a nice way to get quite close to that ideal but there are other ways as well.

However, as said, not all transducers called "ribbons" are real ribbons and they show very different performance.

In 70s appeared first flat tweeters where no transformer was used. They are not ribbon drivers but called isodynamic. The magnetic field was first created with small samarium cobalt bar magnets, which were on front and rear side of the transducer. The magnets, facing each other, create magnetic field across the axis of the transducer and along the direction of the foil. Radiating foil is plastic, for example polyimide, with copper conductors etched on its surface. Front and back plates have narrow slit openings, through which the foil radiates sound. Because the radiation happens through holes, there are usually some response irregularities. Also the flat foil is subject to bending and inherent resonances. Often this type of transducer has low level of high order nonlinearities. Due to large surface area the power handling is somewhat better than that of a pure ribbon. The same area causes drawback in directivity. This type of drivers are currently on the market, but instead of SmCo magnets, neodymium magnets are now used.

As SmCo magnets were expensive, another construction appeared in the same family but used ferrite magnets. It is a relative to a normal dynamic driver magnet circuit. Instead of round air gap it has two wide straight air gaps on both sides of the pole piece. A flat diaphragm is placed in the field; again the conductors are etched to the foil just like in printed boards. This type of transducers offer wide frequency range and less ripple as the radiation happens directly, not via holes. The flat diaphragm usually exhibits similar kinds of high order distortion as said before. Two radiating surfaces cause problems in off-axis response. Often there is a ridge in the centre of the driver, partly isolating the two radiating surfaces from each other.

In early 70s was introduced yet another kind of transducer, where the flat foil is corrugated like bellows in accordion. Magnetic field goes in the front-back direction. Due to the direction of etched conductors and folding of the foil, the folds move horizontally, not front to back, and alternatively move closer to or farther from each other. When moving they squeeze air in and out from the folds. It is said that this gives four to five times more acceleration to the air than a direct radiator. Whether acoustically beneficial or not, this is undoubtedly true, and in this respect it does not differ from a compression driver, where the same effect is achieved with same surface area ratio between diaphragm and throat. Unfortunately the consequences are similar. At higher levels distortion tends to rise and this structure is also subject to the same high order nonlinearities due to bending of the foil, as discussed earlier. It gives the distinctive sound character to this type of drivers. Although the foil surface area is large, it cannot cool effectively because it is folded. When hot, radiation heat is directed to the foil itself and not to outside air, degrading reliability. Another compromise concerns directivity. Larger area is directive at high frequencies; this does not naturally depend on what type of radiator is in question.

In Hifi field there are still other types of flat transducers like Magnepan, which can be called magnetostatic, and the old Apogee, which had large area aluminum foil. I will not go to those in this reply.

From sound quality standpoint the pure ribbon is a superior transducer, while the other flat or folded foil constructions have some inherent audible compromises. Their value depends on design and application; sometimes people want to hear distortion without realizing it. A visit to real concert with only acoustic instruments is highly recommended for a reality check. From purely technical standpoint a properly engineered hard dome tweeter can have extremely low distortions, good power handling, good directivity and it can be very reliable, but its upper cutoff frequency is usually around 22...28 kHz, unless special dome materials are used. Beryllium, some composites and diamond can push the dome upper cutoff frequency to that of pure ribbons.

There are many parameters to be weighed against the requirements of the application and the final choice depends on how different aspects are valued. All real world drivers are some sort of compromises, which is fascinating in the quest for still better fidelity.

GENELEC
Ilpo Martikainen
D.Sc (Tech.) h.c.
Chairman of the Board and founder of Genelec Oy

The Stage Accompany Compact driver is an isodynamic driver where the voice coil is etched on a high temperature plastic foil. As the conductor is long, it does not need a matching transformer. The magnetic field is directed from the pole pieces on the sides to the centerline of the diaphragm. The surface area of SA 8535 diaphragm is quite large.

The on-axis measurements show pretty flat response from 1,5 kHz to about 15 kHz. At higher frequencies the response seems to gradually drop, at least on the sample we have measured. This is easy to equalize if flat response is preferred. Also the distortion products at 90 dB SPL at 1 m are small. 2nd harmonic is between 0,1% and 0,3%. The driver is intended to produce much higher outputs than 90 dB.

Due to width and height of the diaphragm the driver is quite directive, especially vertically: the manufacturer specifies 40 degrees vertically at 2 kHz.

This should be taken into account in the application. Like all line sources this driver produces uniform beam, which is as high as the diaphragm. When moving off the actual beam the response drops but the rate depends on frequency. Vertical coverage can be improved by shortening the output opening with absorptive foam pads on both ends of the diaphragm, but it reduces the output.

Best regards,

GENELEC
Ilpo Martikainen
D.Sc (Tech.) h.c.
Chairman of the Board and founder of Genelec Oy