Fostex FE108-Sol in the BK108 Enclosure — An Update

Some time has passed since the Fostex FE108-Sol drivers replaced the former Sigma models in my back-loaded horn enclosures, and I have had enough time to experiment further with them. As mentioned earlier, removing the notch filters, which were necessary for linearizing the frequency response of the Sigmas, brought a great deal of refinement. I realized how destructively the combination of resistor, capacitor, and inductor in the signal path could act, and looking back now I can say—there is no way back to using them.

Although the frequency response of the FE108-Sol is much more linear, it is still not as flat as I would like. There is still a hump of approximately 3 decibels in the midrange between 800 Hz and 4 kHz, but I have decided not to exchange their impeccable timing and immediacy for a flatter and more accurate tonal balance.

What I learned from this experience is that one cannot have both, at least with this type of speaker, and has to decide whether to enjoy their innate ability to depict events in the most immediate and direct way, revealing the presence of life in the recording, or to preserve a comfort zone by flattening the response, thus preventing the upper midrange from intruding, but also draining away part of the life and atmosphere captured in the recording.

Touching the signal for frequency-linearization purposes will always lead to degradation, and even if it is comparatively small, it will still be detectable. It does not matter whether the correction is performed in the digital or analog domain—any attempt at equalization will leave its traces, however minute they may be. In the digital domain, quantization noise is unavoidable with DSP correction. Any analog correction that routes the signal through additional passive or active electronic components is also prone to introducing noise through phase shifts and energy losses.

In this respect, I would say I am a purist. Although these techniques might be useful and even necessary for certain types of speakers, I doubt that one can truly tame a back-loaded folded horn speaker without causing serious damage to one of its most precious virtues—the ability to reveal the drama embedded in the music.

Of course, folded horn designs have their weaknesses, not only in the amplitude domain (frequency response) but also in the time domain, related to the decay time of signals across the frequency spectrum, as reflected in the so-called decay-time waterfall graphs. Such decay-time-related unevenness is the main reason for the presence of so-called box colorations, which can seriously compromise the fidelity of an audio system.

Apart from the driver specifications, these decay-time-related issues depend entirely on mechanical factors such as enclosure stiffness and resonant behavior, but also on the type of damping material used, its quantity, and its distribution. In this respect, Günter Damde’s decision to double the thickness of the enclosure’s side panels—from the 1.5 centimeters originally conceived by Fostex to 3 centimeters in the Art of Sound version of the BK108—is a positive development.

I also appreciate Damde’s puristic approach regarding damping material, which Fostex engineers leave open for the user to decide according to personal preferences and taste. Damde uses a 1-centimeter-thick self-adhesive damping mat to tackle internal standing waves at the rear side of the compression chamber and along the bottom (horizontal section), as well as the rear (vertical section) side of the horn throat until it reaches the bottom of the cabinet.

I spent quite some time experimenting with different types of damping materials, such as wool and foam, in both the compression chamber and the horn mouth. After all these trials, however, I returned to the solution initially implemented by Günter Damde. The design of these speakers is such that if one decides to add more damping, one may reduce decay-time-related fidelity degradation, but at the expense of losing some of the finest low-level details that contribute greatly to the spectacular, lifelike presence these speakers are capable of delivering.

Therefore, one has a choice: to adjust the speaker’s output according to personal preferences. Especially when foam-based damping, such as Tyrotex Wiffle-Foam, is used in the horn mouth and polyester or wool felt in the compression chamber, these speakers begin to sound much more like ported-box designs and, let’s say, typically Hi-Fi. The bass becomes drier and tighter, the midrange cleaner, and the soundstage appears deeper. But, as I mentioned, this comes at a price—the presentation is robbed of its most precious gem: liveliness.

The phenomenon of liveliness that a record carries is locked within the low-level details and fine nuances of the complex harmonic structure that additional damping tends to diminish. Therefore, regarding the amount of damping, my opinion is that less is actually more.

One also has the opportunity to use sand to additionally damp resonances by pouring it into the pockets located just below the mouth opening. Although such an intervention reduces the internal volume of the cabinet and consequently shifts the harmonic structure upwards, it can further improve clarity, detail resolution, and the sense of weight in the lower registers. I used to have these pockets filled to one-third of their height with sand, but after properly decoupling the cabinets from the floor, there was no longer a need for it.

The advantage of all these adjustment possibilities is that one can tailor the presentation, to a certain extent, according to personal taste. Generally speaking, the Hi-Fi approach (using more damping) benefits a much broader range of recordings and musical styles, allowing even poorly recorded music to remain enjoyable. The less-damped approach, on the other hand, makes bad and overly bright recordings almost unbearable. However, if you are a timbre-oriented listener and have plenty of natural and live recordings in your library, you will benefit from the directness and transparency that convey the atmosphere of the performance in its full glory.

While listening to José Feliciano’s Alive Alive-O! performance, I found myself sitting among the audience in the hall, completely forgetting about any equipment-related matters. There were no speakers, no amplifier, and no digital processor—just the singer, his guitar, and the cheering crowd.

Good acoustic studio recordings, such as Kerstin Blodig’s album Trollsang, although not as spectacular, sound very convincing as well. One still feels connected, but this time to the studio environment. Every detail is presented—the successive strokes of the pick along each individual guitar string, her warm and soothing voice, all providing clues about the exact placement of the microphones. One can almost imagine the recording engineer behind the glass wall, leaning over the mixing console while carrying out his routine.

But don’t get me wrong—these speakers are not studio monitors, and as I mentioned earlier, when insufficiently damped they are not free from colorations. The frequency-response graph reveals a rather unpleasant peak in the upper-bass region around 160 Hz, which, combined with decay-time-related imperfections in the low-frequency region, contributes to specific enclosure-related colorations.

In the case of Kerstin’s album, this makes her guitar sound somewhat unusual in the lower octaves. One gets the impression, especially when the fifth and sixth strings are played, that the guitar somehow has a larger body and a different shape. I can almost imagine a mixture of a Western acoustic guitar and a Persian oud.

Listening to the song Bergtatt with my Beyerdynamic DT 880 Pro headphones confirms this impression. In reality, she uses a standard six-string Western acoustic guitar, which, despite being close-miked in the studio, has all the conditions necessary to sound natural and free from any added hall reverberation. These are exactly the kinds of anomalies that folded back-loaded horn designs tend to expose to some degree.

In the end, these slight deviations in the upper-bass tonal balance do not damage the overall harmonic structure, and the atmosphere of the event remains present with all its palpability. The sense of weight of the instruments, spatial information, dynamic contrast, and timbral nuances are such that one completely loses oneself in the music—something that the DT 880 Pro fails to achieve.

The knock on the guitar’s soundboard is as vivid as if it were happening right between the speakers, and with closed eyes, one can hardly doubt its realism.

Apart from damping, I stumbled upon another important factor that plays a crucial role in the performance quality of these speakers—the mounting of the drivers. I was still using the old Sigma drivers when I decided to mount them using threaded inserts. This became necessary because the frequent mounting and unmounting procedures I performed while testing the influence of different damping materials on the sound had already worn out the screw holes, making it impossible to tighten them properly.

At that time, I did not pay any attention to the sonic consequences of this change. It was simply much more convenient to be able to remove the drivers without risking damage to the fragile threads carved into the multiplex front baffle.

After installing the new FE108-Sol drivers, however, it somehow occurred to me that different tightening torque applied to the bolts produced noticeably different sonic results. This was easier to distinguish because the M4 bolts were much more efficient than wood screws when tightened with the same torque.

It became obvious that lower torque produced better results. With higher torque, the soundstage became smaller and seemed to move several meters backward, dynamic contrast was reduced, and the entire presentation felt somehow restrained. I reached the conclusion that these full-range drivers simply need the freedom to sing.

Considering the fact that introducing threaded inserts creates a much more secure coupling between the driver and the baffle compared to wood screws, I decided to compare both solutions for sonic differences at different torque levels. I rotated the drivers 20 degrees clockwise and screwed them directly into the baffle. Since I was now aware of the possible differences and was actively listening for them, it became apparent that when coupled with wood screws these drivers sounded rather dull.

Although I tried different tightening forces to compensate for the greater resistance caused by the larger thread pitch, it became obvious that when mounted with wood screws they sounded unspectacular and ordinary.

Introducing threaded inserts as a mechanism to reinforce the coupling between driver and baffle resulted in quite spectacular improvements across all aspects of reproduction. As I mentioned, the bolts do not need to be tightened firmly; quite the opposite—they require very low torque so that the gasket between the driver and the baffle can absorb the resonant energy of the driver, preventing this mechanical energy from bouncing back and affecting its performance.

What I think the bolts actually do is to limit the forward movement of the driver in a controlled way, allowing it to move back and forth within a certain restricted range. The important point is that implementing threaded inserts with bolts improves mechanical stability by reducing the tolerance of unwanted forward and backward excursions of the driver.

The original gaskets were already heavily worn, being as thin as a sheet of paper, so I needed replacements to continue the experiment. I decided to try a putty-like material such as Bostik’s Blu Tack, which is appreciated by many as a DIY alternative to standard foam tape solutions. Michael Methe’s latest Audio Note AN-E speaker project is not an exception, and many DIY enthusiasts prefer putty-type materials over conventional neoprene foam.

I ordered a package of Blu Tack, but being impatient, I first tried a regular kneadable putty eraser that I already had available.

It is very similar to Blu Tack, with slightly elastic properties, but as it turned out, it was softer than Bostik’s solution. Even with this type of putty rubber, I achieved rather satisfying results and continued experimenting with different thicknesses. However, because of its softness, it did not allow me to create a gasket layer thicker than approximately half a millimeter.

When the ordered package of Blu Tack finally arrived, I realized that although it might work well with larger drivers (8 inches and above), it was too stiff to be used exclusively with these 4-inch drivers.

I tried several combinations of Blu Tack and putty eraser, and eventually settled on a mixture consisting of equal parts Blu Tack and putty eraser, creating a gasket approximately 0.73 millimeters thick. The results were quite remarkable. Before tightening the bolts, I placed four guitar picks, each 0.73 millimeters thick, between the driver and the baffle to establish an even gasket thickness.

After tightening the drivers, the picks were removed, the excess putty was cleaned away, and the bolts were loosened slightly. It was fascinating to observe how the overall sound changed according to the torque applied to the bolts.

Because the gasket was now made of a sticky material, it was possible to remove the bolts completely. I did this in order to determine their exact contribution, even when they were only loosely tightened. Without them, although the presentation remained quite large, it lost focus and suffered in almost every aspect.

Once I tightened the bolts only very slightly—so that they could not compress the gasket further but merely restrain the forward movement of the driver—the impact on the reproduced music was tremendous.

I was pleased because this very important point—the interface between the driver and the front baffle, where unwanted resonances may occur—had finally been addressed.

After solving the first resonance-prone area, I became aware of another one: the interface between the speakers and the floor. I had ignored this issue for a long time because all my previous attempts to decouple the speakers using spikes had failed miserably.

This failure could be explained by the use of protective discs beneath the spikes. According to my experience, these discs severely compromise the decoupling effect that spikes are intended to provide, especially when used on wooden or laminated floors.

After discovering grub screws in a nearby hardware store, which, despite having conical ends, were not sharp enough to penetrate the floor covering, I bought eight short pieces, each 1 centimeter long, simply to test them under the four corners of the speakers. I fixed them temporarily using small balls of Blu Tack.

Trying them with the song “Vete de mí” from Enzo Pietropaoli’s DOS in Sight album left me stunned. I had not expected such a tremendous improvement in clarity, weight, and separation.

The attacks of the double bass emerged from a black background, expanding beyond the speakers with a power and density I had never experienced before. Eleonora Bianchini’s voice, soft and mellow, was infused with humanity and visceral presence.

Next, I played Vicente Amigo’s “Sevilla” from his album Memoria de los Sentidos, where the knocks on the soundboard of his flamenco guitar and the extremely fast rasgueados were reproduced with a clarity and naturalness I had never heard before.

I then bought 6-centimeter-long M8 grub screws and, using cap nuts, installed them directly into the bottom ends of the cabinet’s side panels. I adjusted them so that the speakers were slightly tilted backward in order to aim the driver axes toward the listening position.

Installing the Fostex FE108-Sol drivers properly and decoupling the enclosures from the floor elevated these speakers to an entirely new level of performance—one I was not prepared for.

The clarity is now such that I can more easily recognize the typical horn colorations in the upper-bass region and the slightly emphasized midrange I mentioned earlier. However, these trade-offs are compensated by their ability to reproduce the atmosphere of naturally recorded performances in a way that very few speakers can achieve.

In the end, owning such speakers can be seen either as a struggle or as a joy—it depends on one’s perspective. For me, it is a fascinating experience and a great opportunity to learn more about audio reproduction by becoming involved in the adjustment and customization of the final link in the chain—the loudspeaker.

Exploring how a particular driver performs under certain conditions and observing the influence that different damping materials have on reproduction is truly an eye-opening (or rather, ear-opening) experience. With most off-the-shelf speakers, one is much more limited in the ability to understand the nature of sound reproduction.

Another aspect is that solving problems through trial and error can greatly contribute to developing criteria for distinguishing between what is true and what is false—something that is rather difficult to define within the world of audio reproduction.

Someone might say: “Hey, it is a matter of taste—there is no objective truth as long as judgments are based on subjective perception!”

I would argue that there is a form of truth that has already been captured at the recording location through microphones and other electronic components during the recording process.

Some people claim that, along with acoustic waves, the atmosphere of an event is somehow captured as well, and I am inclined to believe this. Apart from the intelligible layers of musical content (the music itself: melody, harmony, rhythm, tone, timbre, etc.) and sonic content (recording-related characteristics such as soundstage, imaging, transparency, and so on), there is another layer of information related to the performance itself—the so-called atmosphere.

I believe this layer finds its way into the finest nuances and low-level details that our brains unconsciously process when they are preserved and reproduced.

What I consider to be the main virtue of these speakers is their ability to reveal this subtle layer of information contained within a recording—the atmosphere of the event. Under the right conditions, this can act as a catalyst for connecting the listener and the performer, transforming the act of listening into a deeply engaging and emotionally satisfying experience.

I hope that some of these observations prove helpful.

Happy listening!