Designing loudspeakers is as much art as it is science.
Over the many decades I worked with Infinity founder Arnie Nudell I was always impressed by the combination of a sophisticated measurement system and the deft hand of the designer: knowing when to take a few turns off an inductor or add a 0.1 mFd cap in just the right place.
Once the drivers in a speaker system have been selected it’s time to work with the crossover and here the ways in which designers operate are all over the map. Some use only their instruments while others rarely need them at all. Most I know start as we do in electronic design: a sketch followed by hours in front of the microphone and test screen getting the curve to look as close as possible to the ideal in the designers head. And even that ideal is suspect.
Few designers work to please the measurement microphone’s cry for flatness, but it’s a reasonable starting point. Once the transitions between drivers are smoothed out and the sound of reasonable flatness, it’s time to start playing music and see what’s wrong.
It’s rarely right the first time around.
The new PS Audio AN3 loudspeaker is a 4-way design like Arnie’s reference design: servo controlled amplified subwoofer, internally amplified midbass coupler, ribbon midrange, ribbon tweeters front and back. Before he passed he and I conspired to design new tweeters and midrange based on the Oscar Heil Air Motion Transformer designs known as AMTs. These are ribbons, as in Nudell’s original designs, but they are not flat. AMTs are folded ribbons and as such can move a lot more air with the same precision and low mass as flat ones. Arnie never got to see the AMT tweeter or midrange (the largest in the world). He would have been delighted with both.
Getting Arnie’s reference and the new AN3 system to sound alike has been quite a challenge, not the least of which is due to the fact we’re using different drivers. Our first steps in the design are to measure and compare the two drivers. The next challenge is to couple the tweeter and midrange together in a seamless transition and then on to the midbass coupler.
In my post The Learning Funnel I suggested the easiest way to tackle complex problems is by narrowing them down to their simplest quotient. For the task of getting the tweeter/midrange combination to share the same sonic signature, we (of course) measured the two and compared the results. This can only get you so far. It does not tell you how it sounds. To do that you actually have to listen. But, listen to what? The two full-range systems playing music or just the two drivers?
More to the story tomorrow.