The Analysis Audio Omega Loudspeaker Project
by Dick Olsher (May 2010)
It’s time for something a bit different. This is more than a review article, it’s a colossal attempt to push beyond the boundaries of an ordinary review protocol. In a nutshell, I describe how to thrust the Omega loudspeaker to a higher level of performance in terms of amplification, an improved crossover network, and use of a new “gadget” - the Constant Dispersion Panel.
Although the Omega was reviewed by Wayne Donnelly in the March 2006 issue of Superior Audio (Omega Review), it is worthy of a revisit. I’ve had brief yet enjoyable brushes with the Omega at several shows in the past, but things came to a head during the 2009 CES. As the sun was setting on THE Show Sunday afternoon, Lesley and I settled down to an extended listening session using program material with which we’re intimately familiar (Lesley’s Jazz Me CD). I should add that the system included a pair of Spectron Musician III power amps operated in mono block mode. Lesley, who is even more critical than I about such things, turns to me and says “they’re really good.” !n fact, the sound was to die for: natural, dynamic, and with the tonal weight of the real thing. Also present were Jeff Wells Spectron’s sales manager and Mike Kallelis, US distributor for Audio Analysis. Convincing these two fine gentlemen that I should be given a chance to recreate the magic of the moment in my own listening room wasn’t that difficult. And so a few months later all the pieces fell into place.
Analysis Audio USA was formed about six years ago with Mike Kallelis at the helm to serve as Analysis Audio’s North American distributor. According to Mike, Analysis Audio was established in 1990 with its factory located in Athens, Greece. Its focus has always been on full range planar-ribbon designs because of the conviction that this type of transducer, correctly executed, can reproduce music more faithfully and more naturally than most competing designs.
Mike points out that while planar-ribbon speakers are relatively simple in concept, they are very difficult to produce, which accounts for the limited number of manufacturers past and present. The most well known past manufacturer is undoubtedly Apogee Acoustics. On the surface the Omega may look much like a copy of the Apogee Acoustics Duetta ribbon speaker. But the similarities are a case of form following function. There are in fact significant differences beneath the surface. Both the Omega ribbon and planar woofer use a Kapton membrane as a backing material onto which the aluminum ribbon is glued. What is unique here is that both the bass panel and tweeter ribbons are three-dimensional in extent. In both cases, the substrate is kept taut while pleated foil traces are carefully glued onto the substrate. This type of construction allows the use of the lightest conductive materials while maintaining long-term reliability. Another neat feature is the application of an ultra-thin protective coating to the aluminum ribbon to protect it from oxidation.
Apogee used to glue their woofer panel ribbon directly to a subframe. To the best of his knowledge, Mike believes that the Analysis Audio planar bass panel is the only one ever manufactured with a true suspension. To quote Mike: “The bass drivers are really what makes our speakers what they are. The suspension is made of a carefully selected rubber compound of a specific durometer. Countless hours were spent "listening" to various materials. The choice was somewhat counter intuitive. Other things taken into account during our research were reliability and the ability to withstand ozone.”
After the membrane is joined to the suspension with a special adhesive, the assembly is fixed to the subframe which is then attached to the main frame and adjusted for proper tension. The whole process is said to be done by hand and is obviously very laborious. The payoff is far better dissipation of standing wave energy relative to a directly clamped membrane which actually encourages the production of standing waves, and of course, greatly enhanced excursion capability. The use of lightweight materials, an optimized suspension, and a powerful magnet array are responsible for the impressive bass punch of this speaker.
Integrating the Omega speakers into my medium-sized listening room (14’x19’ foot print with an 8’ to 12’ vaulted ceiling) proved a bit more difficult that I had expected. The wall behind the speakers (rear wall) had been treated with a significant number of acoustic panels, while the front wall is much more reflective. While this semi dead-end/live-end setup has worked well with just about every speaker I’ve had in the house, it appeared to be less than ideal for the Omega. My main complaint had to do with the treble balance; the presentation was much too recessed and polite for my ears. Experimenting with room placement, mainly by pushing the speakers a bit closer to the rear wall helped – but only marginally. In-room frequency response measurements were inconclusive. By that I mean that the near-field measured response was rolled off in the treble - but that is exactly what you would expect from a line-source ribbon. Dispersion is decent in the horizontal plane due to the ribbon’s narrow width, but vertically the treble beams, and more severely so with increasing frequency. Thus, a microphone centered on the ribbon would miss treble output as beaming became more and more pronounced.
Rear-wall dispersion may well be the answer for many dipole radiators. In Las Vegas at THE Show the Omegas sounded their best with large potted plants positioned directly behind them. When the plants were removed, the soundstage lost much of its dimensionality. But how to go about controlling the ribbon’s back wave in my room? I recalled the late Jason Bloom of Apogee Acoustics fame describing his home setup. It consisted of a pair of Apogees situated in front of a large bay window with lots of angled panes of glass. That was the trigger for experimenting with what came to be called the Constant Dispersion Panel (CDP). The idea is to take the rear wall out of the equation when it comes to the treble by presenting the ribbon with a reflecting surface that is also dispersive of treble energy. The Constant Dispersion Panel (CDP) is free standing and about 6" wide and 70.75" long. It leans against the top of the Omega - directly behind the ribbon. My pair of CDP's are made of laminated MDF. Wood or plywood would work as well. There is no fabric cover or any other absorptive material. They look commercial in the picture only because I had my cabinet maker build them to my specs after having experimented with ugly wood planks.
The idea is to provide hard reflective surface of critical width and angle to disperse the treble backwave of the line-source ribbon. The width of the CDP becomes a progressively smaller fraction of a wavelength with decreasing frequency, so that in the lower midrange, sound diffracts around the CDP. With increasing frequency above about 2 kHz the backwave is dispersed along its length in a predictable and uniform manner irrespective of the room's rear wall. Imagine the ribbon being made up of a series of short line segments. In this thought experiment, each segment projected backwards toward the CDP is reflected at an angle of about 23 degrees (from vertical) toward a different spot in the room.
Interestingly, the near-field frequency response is essentially unchanged with the CDP in place. But perceptually the impression of space and rhythmic drive is greatly enhanced when the CDP is in the system. Spaciousness in this case appears to be directly linked to having more uniform treble dispersion in the room. The investment in materials is pretty minimal so by all means do give them a try. They may not be a panacea, but the CDP's should work well in many rooms.
A New Crossover Network
As sold in the US with an external crossover, the Omega is amenable to crossover substitutions. Having direct access to the drivers invites experimentation; for me, it was akin to waving a red flag. I had initially toyed with the idea of experimenting with active crossovers, but a couple of considerations kept me from going in that direction. First, there are practical issues to do with the added expense of a good active crossover, and more amplifiers and cable. Even more important is that off-the-shelf electronic crossovers offer a limited selection of slopes and crossover frequencies. Most audiophiles consider a crossover network as a text book electrical circuit that can be designed with the aid of simple formulas. In practice, the situation is not that simple and one has to take into account each driver’s acoustic slope and the fact that usually drivers possess non-coincident acoustic centers. For example, the Omega’s stock network, shown below, is about as simple as it gets: a first order low-pass for the woofer consisting of a single coil and a single capacitor high-pass for the ribbon.
The text book says that such first-order networks should be connected with in-phase polarity. However, because the tweeter’s low-frequency acoustic rolloff results in about a +90-degree phase shift near the crossover point, an in-phase connection produces a notch in the frequency response. The solution, as you can see from the schematic, is to connect the tweeter with reversed electrical polarity.
The stock network achieves a uniform response through the crossover region. But having measured the individual responses of the drivers it was clear that the response of the woofer panel was exceptionally wide open and that there was considerable overlap between drivers. I assumed that by reducing the overlap I would improve low-level detail resolution, but I didn’t want to stray too far from the spirit of the stock network. A filter’s time domain performance is directly related to its steepness. A gentle rate of descent translates into a better step function response. Thus, I was determined to stick with first and second order electrical filters. I experimented with the value of inductor L1, eventually nudging it in value from 1.5 mH to 2 mH. In view of the woofer’s low impedance, and since this inductor is in series with the woofer’s voice coil, it is critical to use the lowest DC resistance coil type available. In switching from a copper foil type to Madisound’s Sledgehammer steel laminate, the decreased DC resistance (even at 2 mH versus the stock 1.5 mH value) was responsible for over 1 dB gain in output through the upper bass and lower midrange. That was a nice bonus. Next, I wished to eke out a bit more lower midrange from the ribbon (down to about 500 Hz), and at the same time, better protect it against low-frequency excursion. The end result was a second order network – see the new crossover schematic below.
Capacitor C1 was increased in value by a total of 12 µF, by soldering two Hovland Musicaps in parallel with the stock cap complement. The new inductor, L2, is responsible for attenuating bass signal more quickly relative to the stock network. Sonically, the new network is more adept in resolving hall ambiance clues as well as transient decay into a recording’s noise floor. In addition, it is a bit sweeter sounding in the midrange and tonally more robust. In head to head listening test, I realized that there was just no going back to the stock network.
What attracted me initially to the Omega was its big tone, its ability to flesh out the upper bass and lower midrange where most of an orchestra’s acoustic power output resides. “Let’s do the music’s fundamental frequencies right” should be the motto of every speaker designer. For me proper tonal weight is a primary trigger for accepting music reproduction as realistic. Perhaps it rubbed off one of my mentors in this business, J. Gordon Holt, but in any event, it’s difficult for me to accept or appreciate a lean tonal balance. I find it difficult to accept the rest of the package even if it’s laced with wonderful detail and imaging. In audio as in life, it pays to know what’s most important to you. And the Omega delivers the tonal weight I crave. It could even be argued that its’ a bit overdone, that the lower mids are a bit too prominent, but the effect for me is totally believable. There’s also that boxless bass with plenty of pitch definition and excellent extension, well below 30 Hz in my room.
When folks ask me why I’m so enamored of large planar speakers, I try to emphasize two key points. First, dipole radiators couple energy to a room in a manner that more closely approximates a live acoustic instrument. It’s a question of producing the most realistic soundfield at the listening seat. It has been known for many years that well over half the energy at the listening seat is due to reflected energy. A dipole midrange with adequate breathing space appears to provide the most immersive soundstage experience. By contrast, conventional box speakers give me the impression of being on the outside looking in. And unlike most box speakers, planars such as the Omega produce the illusion of live music even when I’m outside the listening room – clear on the other side of the house. That should be proof enough that dipoles sound more natural. The other key ingredient is wave launch. While height information is not encoded in a recording, as microphones simply pickup sound pressure or velocity modulations in point in space, a planar woofer and line source ribbon can synthesize a wave launch that approximates the height and width of the real thing. I’ve never been interested in pinpoint imaging, which has been hailed by some pundits as a desirable attribute. To my ears, tiny focused “balls of sound” are artificial constructs that I’ve yet to experience during a live performance. Size matters. Live instruments do have breadth and height, and large planars do that far better than the competition. A good example of that is piano reproduction. Due to its large soundboard a live piano outputs a lot of acoustic energy over a large surface resulting in a low value of surface loudness density. A planar does justice to the original surface loudness, while conventional boxes seem to squeeze or birth a piano out of an 8 or 10 inch woofer. If such an image sounds painful, it is, and does not work for me. On the other end of the image size spectrum is the trumpet. It sounds piercing because it outputs a large amount of acoustic energy over a small surface area. The Omega also does that well, as the line source ribbon handles much of the trumpet’s midrange and overtone structure.
As much as I like full-range electrostatics, I would be the first to admit that they can’t handle macrodynamics with any kick ass conviction. Even the Apogee planars of old lacked sufficient slam. It’s a different story with the Omega. But there’s a catch… they clearly need a lot of power. So far the best amp (but see late breaking news below) for jump starting the Omega has turned out to be the Spectron Musician III with all the trimmings (Bybee Quantum Purifiers and V-Cap options). As I’ve said before, I'm basically a tube guy but try to stay amplification agnostic, keeping an open mind about what sounds best in a particular application. A single amp works well, but a bridged pair of amps is even better with two kilowatts of headroom! To be honest, getting to love the Spectrons took some time. The Spectron takes a long time to break-in and it also benefits from a long-warm up period. Their sound improved materially after I installed a dedicated 20 amp AC mains circuit to service them, and that’s when I became a believer. The Spectrons generated a spectacular level of dramatic tension, like a coiled spring ready to pop. And when those kick drum whacks do come along they seem to be reproduced effortlessly and without any complaint. Orchestral crescendos evolved linearly and without any hint of compression or obvious distortion products. The Spectrons are equivalent to putting a tiger in your tank. And yet they also have the finesse to caress transients and allow the harmonic envelope of a solo instrument to reach full bloom with realistic colors. They don’t sound like typical solid-state or tube amps. The Spectron’s sound is hard to categorize – it’s both soft and assertive as the music demands it. In my view, it approaches the amplification ideal of being hermaphroditic in terms of textural character. Tube amps tend to chronically soften the sound while solid-state amps tend to emphasize the testosterone aspects of the music. To its credit, the Spectron achieves the Goldilocks middle ground.
It turned out that the Omega is quite comfortable with Class D amplification. Shortly after CES 2010 I was fortunate to spend a few days with the Arion MK 1000 monoblocks ($3,900/pr), a new offering from Arion Audio a sister company of Analysis Audio USA, and I would like to share my first impressions with you. Power rating is 500 watt into 8 Ohm and 1,000 watt into 4 Ohm. Although the Arions could not quite match the Spectron’s sense of dynamic menace, and require a long warmup period (per Mike’s suggestion I left them powered on all the time), I found them to sound cleaner, smoother, and even more detailed. In fact, the Arion’s ability to resolve complex harmonic layers was nothing short of amazing. The harmonic envelope of individual instruments bloomed, ebbed and flowed with remarkable conviction. I expected and got well-defined and punchy bass lines, but I didn’t expect sweet midrange textures and world-class spatial delineation. I found it to be very revealing of components upstream. The Arion loved being mated with Accustic Arts’ new Tube Preamp II, a magical line stage, and I’m sure that there was quite a bit of synergy at play. This should come as no surprise, as I had been advocating the use of tubes ahead of solid-state power amps for many years as a means of coaxing the overall sonic character toward the harmonic color saturation and textural purity of the real thing. Although it was being fed by a reference-class front, at no time did the Arion appear to be a sonic bottleneck. High-energy recordings boogied along suavely and with plenty of verve. The sense of transparency was so spooky that I felt that I could reach into the soundstage and touch someone. Despite its modest looks and price tag, I have to tell you that the Arion MK 1000 is currently my favorite amp for driving the Omega.
I suspect that some of you out there can’t handle the truth. You may be more comfortable within the confines of a conventional HiFi paradigm. Obviously, if you’re in search of pinpoint imaging or a hyped-up presence range, the Omega is not for you. And if your listening space does not allow for at least five feet of breathing space from the rear wall, look somewhere else. Assuming that you qualify and are ready to embrace musical realism, take note: the Omega is a fantastic speaker. For me it represents a must own reference-caliber speaker that will serve me well for years to come. Musically, it is head and shoulders above a host of far more expensive Hi-Fi conventional box speakers.
Woofer: Planar Magnetic, effective area 564.2 sq in
Midrange/Tweeter: Direct Coupled Ribbon, effective area 46.5 sq in
Impedance: 5 Ohm nominal
Frequency Range: 22 – 20,000 Hz
Crossover Frequency: 650 Hz
Recommended Amplifier: 50 - 400W
Dimensions: 66”H x 24”W x 2.4”D
Weight: 99 lbs.
Price: $22,000/pair; $2,400 for the standard outboard crossovers with metalized poly caps, $4,800 for the external reference crossovers with exotic caps.
Analysis Audio USA
385 Forest Hill Way
Mountainside, NJ 07092