The story behind ELAC's innovative Adante speakers
Q&A with legendary home speaker designer Andrew Jones
Andrew Jones is Vice President of Engineering for ELAC America. During his illustrious career in hi-fi, he has designed speakers at KEF, Infinity, Pioneer, and Technical Audio Devices. Many audiophiles consider his $80,000 TAD Reference One to be one of the greatest speakers ever made.
Andrew's recent ELAC designs include the popular Debut and Uni-Fi series, which have received rave reviews for their sound quality and value. Andrew spoke with Crutchfield’s senior home A/V editor about ELAC’s new Adante line.
Crutchfield: ELAC’s Debut and Uni-Fi speaker series have established new benchmarks for sound quality in their price ranges. Do you expect Adante to do the same in its higher, more competitive price range?
Andrew Jones: That was the goal. I’d done some work in the very high end with TAD, and seen that market shrink quite a bit because of the recession.
What happens with a lot of high end companies is they can’t lower their price points because they can’t manufacture a product priced that low in their own facilities. Their overheads are too high. They don’t have the resources to build a larger number of lower-cost speakers, so they go even higher in price.
I noticed a couple years ago at the Rocky Mountain Audio Fest they started an awards ceremony, and the first year I sat there listening to all the categories, and they announced, “And now the category for best loudspeaker over $100,000.” And I’m thinking, “Crikey, that’s crazy! There are enough speakers over $100,000 to make it a competition!”
So I realized that what ELAC could do is take what we’ve been doing with the really affordable but good sounding stuff, and move upward to entry-level high end sound quality, but at a much more affordable price.
So for the tower, the goal was $5,000 a pair. Could we get something that looks high end, and sounds high end, and competes with, let’s say $10,000 speakers, but for $5,000?
And for the stand-mounts, the goal was $2,500 a pair.
ELAC's Adante AF-61 floor-standing speaker and AS-61 stand-mount speaker.
Crutchfield: Could you talk a bit about some of the innovative design in Adante? For one thing, it uses a multi-chamber cabinet where there isn’t really an external port.
Andrew Jones: Yes, so I knew it was going to be a 3-way concentric design because I like doing concentric speakers. I wanted to do a better midrange driver, and an especially effective tweeter. The tweeter in the Uni-Fi was good, but I knew I could do better. So that was one goal.
The second goal had to do with the bass. I wasn’t initially thinking, “Oh, let’s do a completely different bass loading technique.” It was more like, with the ported boxes I’d done, which were all rear ported, we’d always had some questions, like, “How close do you put them to the wall to increase the bass?” Although technically, that’s not really as much of an issue as people might imagine.
On the other hand, why invite questions? So, it got me thinking. I wasn’t going to do a sealed box because I lose too much capability with a sealed box. So I then started thinking about something I’d done nearly twenty five years ago, which was basically a “coupled-cavity” type speaker system.
I thought, “What if instead of having a port on the front, I put a passive radiator so everything communicates through the passive radiator?” That way I can do something that’s different, might solve all my problems, and give me some performance advantages
Afterwards, it became “What was I thinking?” [laughs] It took much longer to develop than I thought. It’s a very complex system, and I ended up spending a long time getting it to work properly.
Cut-away view of the AS-61's Interport-Coupled Cavity enclosure. The internally-mounted 6-1/2" woofer and dual ports communicate bass energy to the front chamber and its 8" passive radiator.
Crutchfield: I can imagine! It’s radically different from anything I’ve ever seen.
You know, here at Crutchfield, we do a lot with both car audio and home audio. On the car stereo side we’ve had bandpass subwoofer boxes for years, where you have a sealed box coupled to a ported box. And those designs can deliver tremendous bass slam at certain frequencies, but not so much musical accuracy, at least not the ones I’ve heard.
Andrew Jones: Well, the coupled-cavity design is actually a fairly old idea. I can remember right before I joined KEF, I attended a lecture on coupled-cavity systems by Laurie Fincham [KEF’s then technical director].
Then when Laurie invited me to join KEF, they were working on the model 104/2, which was the first implementation of coupled-cavity in a hi-fi speaker. KEF didn’t invent coupled-cavity, but what they did was invent the mathematics for designing it.
I came up with a new version of coupled-cavity which I call “interport.” The way I like to describe it is instead of putting an extra box over the front of a sealed-box loudspeaker, you put an extra box over the front of a ported-box loudspeaker.
Andrew Jones, ELAC America's Vice President of Engineering.
So, both the driver and the port drive into the front chamber, which is then either ported, or in this case, taken further, by using a passive radiator.
Crutchfield: I know that computer modeling plays a big part in speaker design, but Adante is so different, is there even is a modeling program that applies to it?
Andrew Jones: Well, I had to develop a new model for it. Although a passive radiator sort of works like a port, it’s not exactly the same. You have to take into account the stiffness of the suspension of the passive radiator.
Anytime you want to change the frequency response, it’s not like simply changing the crossover. You’ve got to physically change the box and the driver and everything else, so we ended up building one box, analyzing it, then trying to fit a model to it to see if the model matches what you actually built.
So you build a box, fit a model to it, see that the model gives you the same result as the box did, then you use that model to calculate something else.
You make some changes in the model – say, to the box size or the mass of the passive radiator. Then you build that and see whether the model now accurately predicts something new. You build it and you check it. Did it turn out like the model said it would? If it didn’t, then you just got lucky the first time. So you have to keep going round like that.
Then you’ve got to put a crossover on it, and crossovers interact with the impedance of the speaker. So how do you make the crossover work with that impedance curve? Because Adante has a very complex impedance curve.
A sealed box has a single resonance spike in the impedance curve. A ported box has a double peak. Well, a coupled-cavity system like this one has a triple impedance peak. So you’re now trying to get a crossover network to work with that, and it’s not easy.
So I thought, “How about if I also add the crossover components into my model?” Now I can optimize everything together. So instead of optimizing a cabinet and driver, then trying to come up with a crossover network to work with that, I can optimize every aspect and have more of a guaranteed result.
Of course, that made my model even more complex, but hopefully once I’d finished I’d have both a practical network and a practical box and driver. I also wanted a nice, smooth impedance curve that didn’t shoot down to 2 ohms or something silly.
That all took a lot of work. At one point I thought, “This should have been a research project rather than a development project!” But eventually I got everything working well enough that I could move forward with building a practical speaker and start listening to it to see how I could get the best performance from it. That took quite a while, too.
Crutchfield: You mentioned earlier that you were able to optimize the system with the enclosure, drivers, and crossover all calculated into the model. The Adante crossover isn’t a run-of-the-mill design, is it?
Andrew Jones: It’s always the combination of the electrical crossover network ahead of the driver, and the driver response, combined, that tells you how they will act together. Typically, a crossover isn’t just blending the speakers together. It’s also trying to cut out all the nasties in the driver and the out-of-band response.
The Adante's crossover system combines a high-quality 1st-order electrical network (pictured) with the second-order "acoustic" filter created by the multi-chamber cabinet.
When I put a box with a passive radiator over the front of a ported-box loudspeaker, that cavity combined with the mass of the port or passive radiator is a second-order acoustic low-pass filter. The speaker already rolls off at low frequencies as all speakers do. But now it’s forced to roll off at my chosen crossover frequency – in this case 200Hz – but it’s a second-order acoustic filter.
If I want a third-order response overall, then we only need to add a first-order electrical crossover. And there’s an advantage to splitting the task between the passive network and the acoustic filter.
Imagine I’ve got a regular woofer and I put a third-order electrical network on it. Well, in a perfect system it’s doing the same job. But what it’s doing is filtering the input signal to the driver, not the output from it, and that’s a subtle but important distinction because all drivers have some amount of distortion.
Now, Adante uses internal ports, but all ports have some form of noise. If I’m only filtering the input signal then I’m not filtering the noise of the port. And I’m not filtering some of the distortion components.
Suppose I put a 100Hz tone into the woofer and it’s got some distortion – the third harmonic is at 300Hz. That comes straight out.
However, if I put a second-order acoustic filter after the driver, that third-order harmonic is attenuated, so I’ve effectively reduced the harmonic distortion of the speaker. I’ve also filtered off some of the chirping noises of the port. So I’ve got an advantage by putting the filter after rather than in front of the driver.
Crutchfield: That is fascinating, Andrew. Since the Adante design ended up being more complex and challenging than you expected, are you happy with the results?
Andrew Jones: Yes, I am! You know, I always know I can do even better next time, but certainly the impressions people have been giving at all the shows – of course it’s only just starting to be reviewed – the impressions are always, “This is very truthful bass.”
Crutchfield: You mentioned that one of the goals for this series was to develop a much better tweeter. Can you provide a few details?
Andrew Jones: Certainly. When you’re developing a concentric driver, you’re trying to get a tweeter that will fit down the middle of the voice coil, right? You’ve got a space limitation, so you’ve got to design within that, which is why the Uni-Q that we developed at KEF only came about once neodymium magnets were developed, because they were strong enough that you could make them small enough to fit in the voice coil.
The ELAC Uni-Fi midrange used an inch-and-a-half voice coil, and that gave me enough room to engineer a tweeter. So my goal in developing the tweeter for the Adante was to extend the bandwidth to go both higher and lower.
Adante's concentric driver places a 1" soft dome tweeter in the voice coil of the 5-1/4" midrange cone.
I wanted it to still be a soft dome because although I am looking to develop metal domes, that’s still a research project. In fact, in the early days of Adante, I thought maybe I could develop a hard dome tweeter. But on top of doing the bass technique it was going to be a lot more development time. So I decided, let’s just go back to a better form of soft dome. A good soft dome is better than a poor metal dome.
Crutchfield: I still generally prefer soft dome tweeters.
Andrew Jones: There’s a particular type of soft dome that started to appear a few years ago. Unlike with typical soft domes that use a single surround, this one used two concentric surrounds, with the voice coil in the apex between them. That gave very extended high-frequency response, and better low-frequency response, too.
But when I investigated that design, I ran into a problem. It increased the tweeter’s overall diameter so that I couldn’t get it to fit in an inch-and-a-half voice coil. So, in order to develop that type of tweeter, I had to engineer a new midrange driver with a 2” voice coil instead of an-inch-and-a-half. And because that then lost me some radiating area, I had to make it a 5” cone instead of a 4” cone.
So that’s what it became – a 5” cone with a 2” voice coil. It gave me room to engineer a wide roll surround soft dome, and now I had response that went out to about 35K. You can argue about whether or not we need that much bandwidth, but traditional soft dome tweeters actually rolled off before 20K. Some of them could be between 3 and 6dB down at 20K. Metal domes weren’t like that, but they did suffer from resonances at 22K. By engineering a tweeter that goes out to 35 or 40K, it means it doesn’t roll off by 20K.
So this tweeter combines a new soft dome material with the roll surround. It goes as high as I need it to with no roll-off by 20K, and it also goes down lower so the resonance frequency is somewhere around 1K. That means it’s easier to cross it over at a lower frequency, so I crossed it over around 2K. The new tweeter design necessitated a new midrange design, so everything changed.
Crutchfield: Earlier, you said that by losing the external port you could place the speaker closer to the wall, if you wanted to. Would you say in general that the Adante line is fairly insensitive to placement, and also fairly easy to drive? I noticed they are 4-ohm speakers.
[Editor's note: At the time of the interview, the nominal impedance for the AS-61 was listed as 4 ohms at ELAC's website and in the AS-61 owner's manual. We later learned that the impedance was changed to 6 ohms late in the speaker's development.]
Andrew Jones: They are, but they’re unusual 4-ohms. The problem is, what do we mean when we say a speaker has a certain impedance?
The international standard is very simple. If you say a speaker is 8 ohms nominal, it must not have the minimum impedance lower than 80% of that nominal impedance. If we say it’s an 8-ohm speaker, it cannot drop below 6.4 ohms.
That’s the standard. Now you’re not bound by law to follow the standard, but speaker companies all got together to create a standard, so why wouldn’t you follow it?
Well, you can cheat and not follow the standard. Create a drop in impedance, which affects the speaker’s sensitivity. A speaker’s sensitivity is measured by driving it from a power amplifier at 2.83 volts – it will generate so much sound pressure.
Now if you lower the impedance you can get it to play louder under most circumstances because you’re actually drawing more current and hence more power. And that’s the dirty secret, especially when you switch speakers the old-fashioned way in a store demo.
So, if we say minimum impedance should be 80%, then for a 4-ohm speaker it should be 3.2 ohms. But at which frequency does it reach its minimum?
If it’s where there isn’t much music power you wouldn’t really care. But for other frequency ranges where there’s a lot of music power, you would care.
Secondly, if it’s always at 3.2 ohms then on average it’s going to draw a lot more power than if it’s only at 3.2 ohms over a relatively narrow frequency range.
So what I do is, I’m totally honest with my minimum impedance. If I say it’s a nominal of 4 ohms it will never drop below 3.2 ohms.
So for the stand-mount it’s actually a 4-ohm system, but only over one small frequency range it’s 3.4 ohms. Everywhere else it’s much higher than that. So it’s an easy 4-ohm load.
The towers are actually a higher-impedance and easier load than the stand-mounts. I’m no longer fighting with a smaller box to get the best performance I can out of it. And it was a slightly easier task when I’ve got three woofers.
Crutchfield: Right. So do you feel like a good quality receiver should be able to drive even the Adante stand-mounts?
Andrew Jones: Well, if we’re talking receivers, we’re really talking home theater, and with a home theater you’re going to have a subwoofer. Even with a speaker with extended bass like the stand-mounts – you know, the stand-mount and the tower both have about the same 3dB point. But where the stand-mount will run out of steam trying to produce high levels of the lowest frequencies, the floor-stander won’t. It can do that all day long.
Even so, with the kind of low-frequency signals you get in a home theater setup, I say get a subwoofer, even with the tower speakers. A subwoofer relieves the amplifier of a lot of power that would be needed to drive low-frequency signals, leaving more available for mid-bass, midrange and treble.
So in that sense you could say, “Yeah, a receiver because it’s going to be used with a subwoofer – it should be fine.” But you still want a receiver with good capability. I’m trying to design high-end speakers at affordable prices.
The difficulty is if you then think, “Oh, it’s a cheap speaker. I can use cheaper equipment.” You still have to be sensible. At least buy a receiver that’s developed with the same ethos that I’ve applied to the speaker – that is, a high capability receiver, and not just a very low-cost model.
Crutchfield: Is there anything else you’d like to say about Adante?
Andrew Jones: No, I think I’ve got through most of what I wanted to say, other than, of course, with Debut and Uni-Fi, the box construction is okay, and the finish is okay. But it’s somewhat compromised in order to give the best sound quality.
So, moving into higher-end price points for Adante, we needed it to look good. And so there is more money spent on beautifully painted cabinets.
We’ve got the aluminum baffle, which is machined both to add stiffness and dampening to the cabinets, and to act as sort of a waveguide for the concentric driver. But also to make it look nicer.
I’ve had some people who’ve already received these who were TAD customers of mine from the past. They were looking for a lower-cost second system, and they’ve come back to me and said, “We spent $80,000 on TAD speakers. We can’t believe what you’re giving us for $5,000 in both construction and sound.”
Crutchfield: Well, we were pretty amazed when we sawed one of the prototype stand-mounts in half – that’s an incredibly well built cabinet!
Andrew Jones: Yeah. And particularly with the floor-stander, there are two chambers for each woofer system, so that means six chambers plus the midrange chamber. You’ve got the box internally divided into seven chambers, so all of that greatly increases the cabinet quality.
Crutchfield: Andrew, thank you so much for taking the time to talk with me today. Good luck.