Jim Thiel: A Coherent Source Page 2
Coincidentally, at about the same time [that I was thinking about time coherence], other companies were also starting to think about the problems caused by the differences of arrival time. We came up with the approach of mounting the drivers on a sloping baffle, which did not introduce the problems caused by the stepped baffles that were being used in some other designs. We introduced our Model 03, which achieved a lack of phase distortion, in 1978, and we still use this approach in all of our products.
Atkinson: How do the Thiel "Coherent Source'' speakers differ in design philosophy from the original 03?
Thiel: In terms of the basics, there are really no differences. All of the products use a sloping baffle to align the drivers. All of the products use a first-order crossover system to achieve a lack of phase distortion, and all of the products have vertically aligned drivers so that, if a listener was off-axis, the relative distance to each of the drivers did not change appreciably. That was kind of unusual back in the '70s.
But all of our products have incorporated incremental advances in various aspects of their design: improvements in the quality of the crossover components, improvements in the drivers, improvement in the cabinet construction. The Model 04, for example, was the first unit where we went a little bit further to reduce cabinet-edge diffraction. The Model 03A went even further in that regard, and we started using higher-quality crossover components. And when the CS series came out, the CS3 was the first model where we used a cast driver chassis and a much thicker cabinet baffle. It was also the first product, I believe, where we used polypropylene capacitors in the crossover network. When we started development of the CS3 product, we originally referred to it as the "03B''—it was the third generation of our original Model 03. Similarly, our current CS3.6 is the sixth generation of that 10", three-way, floorstanding, phase-coherent, sloping-baffle design that we first introduced in 1978.
Atkinson: Is a first-order crossover as simple as just putting a series capacitor in the tweeter feed and a series inductor in the woofer feed?
Thiel: Unfortunately, no. If you look in the textbooks, that's what they may show you as a first-order crossover system. And that would be all that was required if the drivers had uniform frequency response for octaves beyond the nominal crossover frequencies. If the drivers had uniform impedance characteristics for octaves outside their nominal bandpass. And if cabinet diffraction effects were not significant. But in the real world, none of those conditions are true and all of those factors have to be taken into account. So in practice, the first-order crossover system can actually end up being more complicated than any other type because we have to compensate for and correct all the anomalies in amplitude and phase that exist through a much wider bandwidth than if the same driver were to be used in a more normal, high-slope crossover system.
Atkinson: Through how many octaves to either side of its passband does a drive-unit have to be flat to make it good for use in first-order systems?
Thiel: At least two.
Atkinson: So for a woofer that you're going to cross over at 3kHz, that means it would actually have to be flat to 12kHz.
Thiel: Right. And sometimes some compromises have to be made. Two-way systems are especially difficult to implement with first-order crossover systems. One of the things that becomes important is that the woofer-to-tweeter crossover point be as low as practical so that you can ease the requirement on the upper limit of the woofer. You need a tweeter that has very high output capability.
Atkinson: Like all new speaker companies, you were using OEM drive-units at first. However, you quickly got into partnerships with your suppliers where you started to specify exactly what you wanted.
Thiel: Yes. Early on, we started making small but important changes to off-the-shelf drivers. However, it's very easy for the driver manufacturer to change the voice-coil, say, or the stiffness of the suspension so that the speaker can be optimized for use in your system with your cabinet and your bass-tuning parameters. We started getting more and more into designing our own diaphragm profiles, tooling our own rubber surround geometries, and making the drivers more and more specific to our systems.