What is important for DIY?
Dr. Joseph d'Appolito Dr. Joseph D'Appolito,
Chief Engineer at Snell Acoustics,
consultant for many manufacturers
one of the most established speaker developer
Author of e.g. Testing Loudspeakers, 1998
many DIY designs:
(no longer available)

Interview to the principles of speaker development,
Peter Strassacker asked Joe d'Appolito (5/2004).

Joe, we want to give our DIY-customers the best information, so they have the chance to build very good kits, which they like 10 or perhaps 20 years. What is your advice to them?

Yes we have gone well beyond the "cook book" phase of loudspeaker design. We now have a wide range of measurement and design tools that allow us to make very accurate speakers.

The other side of this question is that both the general public and the DIY types demand the better sound, but do not know how to get it. For the DIY type, professionally designed kits are the only way to get today's level of performance.

As I stated earlier, in modern design we cannot just use some formulas in a circuit design textbook. We must accurately account for the impedance and frequency response of each driver and determine the complimentary crossover for that driver. There are also questions on driver types, multiple driver integration, controlling diffraction, controlling system impedance, etc.

Yes, the driver, the cabinet and the crossover must fit and the crossover
is just good for one combination of drivers. What are your ideas
on driver types, multiple driver integration, diffraction etc.

Driver types: The most common driver is the dynamic driver. These include the dome tweeter, dome and cone midranges and cone woofers driver by a motor consisting of a voice coil and magnetic circuit structure. This class of driver has been around for over 80 years and during that time it has matured into an excellent transducer. These drivers are relatively easy to integrate into multiple driver systems and are therefore most useful in developing successful DIY projects.

More exotic drivers like ribbon tweeters and ribbon midrange drivers are appealing to DIY experimentors, but their polar response patterns are quite different from their cone counterparts, making it more difficult to integrate them with cone type drivers in multiway systems.

Multiple driver integration: As more drivers are added to a loudspeaker design getting them all to work together properly becomes increasingly difficult. Cut and try methods cannot provide the level of performance we demand of today's loudspeaker. In modern design we cannot just use some formulas in a circuit design textbook. We must accurately account for the impedance, frequency response and acoustic position of each driver to determine the complimentary crossover for that driver. This requires sophisticated measurement equipment and loudspeaker modeling and optimization software. For the DIY type, professionally designed kits are the only way to achieve the level of performance demanded today.

Having said that how complex should a multi-way full range system be? Personally I feel that today's dynamic drivers have reached the point where you can get all the performance you need with a three-way system. As one goes beyond a 3-way speaker getting all the drivers to integrate correctly becomes more problematic. A 4-way, for example, usually forces us to crossover in the very critical midrange. Getting a seamless transition in this region is very difficult.

Diffraction: Diffraction is much talked about and little understood. In home audio application the diffraction we are most concerned with is that occurring when sound from a driver reaches a baffle edge, where a secondary emission of sound occurs. There are two issues here: 1) How do we control diffraction? and 2) How important is diffraction to the sonic signature of a loudspeaker? The first is an engineering question, the second is a matter of opinion.

1) First a myth. Rounding corners eliminates diffraction. Diffraction, like everything in acoustics, is wavelength dependent. With vertically oriented speakers, diffraction is mostly off the vertical baffle edges. If the rounded edge is to provide a smooth transition for the sound wave off the baffle, the radius of the edge must be on the order of a wavelength. A one inch (25.4mm) radius corresponds to a frequency of 13.5kHz. So even this generous edge has no effect on the critical midrange and presence range. Diffraction is also angle dependent. Its effect diminishes rapidly as you move off axis. So diffraction has little effect on the loudspeaker power response. Since we use the power response to judge a speaker's frequency balance, diffraction, unless it is extreme, has little effect on perceived frequency response.

2) What about imaging. Here we get subjective. We use first arrival response to judge source location. This happens in the first 300-400 microseconds. Diffraction is a response delayed by the time it takes to get from the driver to the baffle edge. This is also on the order of 300 microseconds. So the diffracted wave arrives some 300 microseconds after the first arrival if you are directly on axis. However, the diffracted sound level is several dB down from the first arrival (Diffraction is well designed speakers typically adds no more than 1dB of ripple to the on-axis response.). In my experience, speakers with this level of diffraction still image extremely well.

Thank you Joe.