Danish biochemist Troels Gravesen has received a lot of well-deserved attention for his outstanding DIY loudspeaker projects. Troels' projects are always well documented, use premium drivers and have beautiful enclosure designs, giving aspiring loudspeaker builders a roadmap toward building status designs that would retail for exponentially more money if they were commercial loudspeakers. His site also features projects of his that others have built, giving us a chance to see different interpretations of his plans. Jantzen Audio sells kits that Troels has designed using Parts Express enclosures, for those whose carpentry skills may be suspect.

 

 

 

 

AV Enthusiast: How long have you been designing loudspeakers?

Troels Gravesen: Ever since I was a boy. I’m 55 now, and well, designing and designing. Like most other diy people I did it by ear and calculation and most of the times, it really wasn’t very good. When PC based measuring equipment became available at affordable prices, I jumped onboard and a few years later started the website to share my experiences.

 

AVE: Was loudspeaker design a self-taught skill or did you have a mentor?

TG: During high-school in the late Sixties and early Seventies, I bought the few books available on the subject: “Loudspeakers” by Gilbert Briggs (Wharfedale) and some German books too. Sometimes it took months to get the books; letters had to be written to publishers, snail mail, transfer of money was a slow and troublesome. Today we can get anything from everywhere at a click on the computer. Briggs’ book on loudspeakers was probably my biggest inspiration and I still read this book from time to time.

 

40 years ago most of what is needed to make good sound was well known. Ever since, it’s been fine-tuning and evolution. The design of crossovers may be an exception. We are much better equipped today to design proper crossovers. With regards to the design of drivers, we can find vintage drivers that easily outperform many modern drivers. I’m buying vintage drivers from time to time – to learn what they did back then – and to compare to modern drivers. Like now, you can find a lot of very poor performing vintage drivers, but sometimes you find pearls, from which there’s a lesson to be learned.

 

AVE: What motivated you to start designing loudspeakers? Was there a loudspeaker you wanted but couldn't afford, or did you just feel that you could do it better? AVE: Your work has a lot of fans. Whose work are you a fan of? Is there a particular designer or loudspeaker that you particularly admire?  AVE: The focus of most of your designs is set squarely on two channel listening. I've noticed that a lot of your designs are relatively low sensitivity. I understand that to get low end extension we have to give up some sensitivity. Are there other sonic compromises that you feel high efficiency loudspeakers have to make that you don't want in your designs? 

TG: I almost never watch videos and surround sound isn’t exactly my cup of tea, thus no designs aimed at this. I’ve done a few centre channel speakers for private customers.

 

I’ve modified my television to connect to external speakers and I have some old high-efficiency Vifa drivers for standard two-channel sound. For new films we go to the cinema and I’m mostly shocked by how good the midrange can be from a proper cinema set-up.

 

I have a fairly small listening room, 25 square metres, and my wife has to live there too, thus a wide range of small designs. If we want at least some bass from small speakers, we have to sacrifice efficiency. We can’t fool nature here. Many commercial designs are often overrated by several decibels not to scare people off.

 

All newer designs on my website are shown from 2.8 volts input at 1 metre distance. As a diy’er, I have no interest in lying about efficiency, thus my designs may appear lower than average compared to commercial designs. We have examples from John Atkinson (Stereophile) on commercial speakers claiming 98 dB sensitivity, which turned out to be more like 92.5 dB. Quite a difference!

 

We almost never see actual measurements on sensitivity from commercial designs and to some extent I can understand this. You need thorough experience in the interpretation of measuring performance to evaluate speaker performance. And even so, this tells very little about how a speaker actually sound. People may start thinking about a small bump here and a small dip there and anything that might leave your potential customer in doubt is commercially bad. Then better tell that this is “the best speaker in the world”. Period.

 

For the last couple of years I’ve spent more and more time on high-efficiency speakers and have a whole section devoted to this.  From this work I have launched several designs – mostly studies that don’t necessarily make plug-and-play design for diy’ers. My articles are learning exercises, trying to find out what the 3 watt SET people are so thrilled about.

 

As said, if we want high efficiency plus low-end extension, we have to give in on size. We need big speakers, sometimes even really big speakers. Commercially we are already seeing designs with 8-10” drivers in fairly small cabs claiming 95-96 dB sensitivity. If we’re talking 40-60 litre volume, we can be absolutely sure these speakers won’t say much below 60 Hz. It’ can’t be done. A German producer of high-efficiency speakers has co-launched a tiny little box meant to go in between pre-amp and power-amp – and oh yes, we have bass equalization to overcome the lack of bass. Do we miss the tone controls of our old pre-amps? Indeed we do when we claim 95 dB and a 40 Hz bass extension from small cabinets. And even tone-controls won’t make 40 Hz from a design that cannot reproduce 40 Hz to any significant level. What we do is lifting the 50-100 Hz region and I don’t say this is bad at all. I don’t think there’s anything wrong in equalizing a sloped bass response flat down to 50 Hz if this is the potential of the design. Most bass is in this region and from personal experience with 8” Supravox drivers, having Fs = 65 Hz, we can get surprisingly good bass – in huge cabs!

 

What I have tried in my JA8008 design is to combine some of the features from both worlds. Modern, well made high-efficiency drivers plus proper crossover management. The latter feature leaves out the “single cab people” as target customers, but these guys are hard to please anyway. Some SET people believe you should run one single driver and sometimes only have a supertweeter handling the upper octave run from a single capacitor. It’s an apparent non-compromise approach, which to my ears implies a lot of compromises as very few drivers can handle e.g. 50-8,000 Hz range properly. There are a few designs that come close as long as you don’t push them too hard, e.g. Supravox and presumably the French PHY-PH drivers. The latter I haven’t heard.

 

I’ve heard non-whizzer Lowthers run this way with noticeable distortion in the treble range. The extremely thin paper cones simply break-up as soon as you crank up the volume just a little bit. But high-efficiency speakers are certainly not only meant for low-wattage valve amp enthusiasts. Any decent sounding solid state amp may benefit from not having to deliver loads of amperes into in-efficient bass drivers.

 

What we get from high-efficiency speakers is the speed and transient attack we simply cannot get from heavy cones. No matter how many watts we put behind, we cannot accelerate a heavy cone as fast as a lightweight cone. Simple physics.

 

What we can have from high-efficiency speakers is balanced tonality, proper phase tracking between drivers in the crossover region and we may be able to address overall power response much better than we could previously. There’s nothing new in this. Studio technicians have made speakers like this for their own application for decades. What we’re seeing now is the attempt to sell this concept for home audio with one little exception: Size. We badly need to educate consumers on the relation between size, efficiency and bass extension.

 

TG: Transparency is a prime goal, that’s for sure. If transparency is good, it tells a lot of things have gone well in the design. Tonal balance is good, phase performance is good, the driver cones may break-up very little and not smear details, etc.

 

We may have all these qualities right and still think a speaker sounds dull. There’s more to good sound than “pin-point” imaging. I very much enjoy a speaker like my JBL L26-3-way reconstruction. Not particularly strong in pin-point imaging, but it’s an “easy to listen to” speaker. The 10 inch bass driver handling most of the midrange is probably one reason. Overall sensitivity is around 90 dB and it’s an easy load on the amplifier due to the acoustic vent used. It doesn’t go deep but the bass beats any single 6-7” design I have made.

 

I can’t say I’m targeting a particular sound, as “sound” as such can hardly be described. Our vocabulary is poorly suited to describe sound, where I think we are much better in describing e.g. taste and smell. I’ve read about native tribes living in particular humid environments describing “colour” as grades of moisture or grades of “wet-ness”. The inuits have numerous words to describe the condition of snow, but despite having had loudspeakers around for more than 100 years, I still think we’re poor in describing sound.

 

We use words stolen from other areas, e.g. smooth, silky, etc. I’ve read about the reproduction of violins from a particular speaker being “silky”. Well, if you stand right next to a violin, we can hardly describe the sound as being “silky”. If this loudspeaker makes violins sound silky there’s probably something wrong about it. I’ve recently read the term “pillowy”about treble performance. Now that’s a new one. Clearly shows how much trouble we have in describing sound.

 

What I try to target in my final voicing of my speakers is a proper balance of basic notes and upper tones. I often use the BBC dip to render this balance and it may to some ears be a slightly laid back voicing. Many commercial designs balance the treble 1-2 dB above average level, because treble sells. If you go into a shop and swap between a range of speakers, most people will pick the most bright sounding speaker. A bit like pushing the loudness button on an old amplifier. More bass and more treble. You can’t stand it in the long run, but it sells.

AVE: Obviously drivers have gotten better as time has gone by, but when you get right down to it, the technology of loudspeaker design isn't significantly different than it was 20, 30, even 40 years ago. What aspect of design or improvement in materials have we made the most progress in?

TG: Hard-cone drivers is certainly an area where things have improved considerably. Ceramic cones didn’t exist back then. Magnesium cones neither. These drivers can deliver uncoloured sound that can hardly be compared to any old drivers. Still they lack sensitivity and we have to look further for similar performing materials, preferably delivering some 3-6 dB increased sensitivity.

 

Paper cone drivers appear to get a renaissance and new matrix materials are seen every year. The problem here is the amount of resources needed to make radical improvements. Making a new paper cone for an 8” might easily include optimising a wide range of parameters, producing 100+ different drivers for sonic evaluation. Few have the resources to do this. Usually the cone producers will try out different things from time to time and go to the actual driver producers and ask them to try this or that.

 

Thank you Troels, for sharing your time with us.

 

If you read through this interview, you obviously care about audio. I urge all of you to pay Troels a visit at his site.  If you have played with the idea of constructing a DIY loudspeaker design, I can't think of a better place to send you for plans that are sure to please both your eye and ear. Even if you hold no such ambitions, go take a look at his projects and read the notes. I guarantee you will leave with an enhanced appreciation of both the science and craft of loudspeaker design.