The real key to this project is in the unique crossover topology that we have been working on and trying to perfect for over a decade now.
One of the problems with 3-way parallel crossovers is that there needs to be a very large capacitor in series with the midrange driver to form the high pass filter, which gently rolls off the lows. Such a large value series capacitor in series with the critical midrange driver causes some unavoidable issues. Amongst these are a large phase shift, distortion, and it also dampens dynamics. Then, to increase the high pass filter slope from 1st order to second order (12dB/octave) – you then must add the appropriate inductor in parallel after this capacitor. This then takes lows that should not reach the driver directly to ground.
To filter out the highs that are meant to go the tweeter, you then also add a low pass filter, (a series inductor with a parallel capacitor) – thus completing what is known as a bandpass filter for the midrange driver. The tweeter then uses a separate high pass filter while the woofer uses a low pass filter, and then these 3 filters are wired in parallel. This is the standard for crossover designs for 3-way speakers. It is a tried-and-true design, relatively simple to design and when done right, provides good results.
I have had an idea in my head for a different 3-way crossover topology for what must be 25 years now and we started work on this back when we first developed our towers. We made several samples, and the results were good, not great – and not better than the more simplistic and common crossover design we eventually went with.
With help from advanced computer modeling and hundreds of hours of NFS testing, we have finally been successful in developing this different crossover topology, that avoids the negatives of a large series capacitor as described previously while also routing much of those low frequencies that would normally go to ground, to the bass woofers.
The results are amazing, driver integration and overall directivity is about as good as it gets for a passive loudspeaker. Coherency is on par with the best 2-way speakers, yet this is a true 3-way. Dynamics are astounding as is bass response as this new crossover design mitigates those losses as I outlined above. Because phase response is more accurate, directivity is improved and this results in remarkable spaciousness, but not overly so such that center imaging is still focused and not dispersed.
We lose a small bit of overall efficiency, but this is offset by presenting an easier impedance load to the amplifier, so the amp uses less current and as such, has more headroom available.
In simple terms, this new crossover brings out the maximum potential of each transducer, sums them ideally and the results are clearly and instantly audible when it comes to dynamics, spaciousness and imaging.
For many reasons I won’t be publishing details on the circuits themselves, but it functions differently than a 3-way parallel crossover and the results have proven to be better than I had expected. I am honestly not sure if it would have been possible to properly design this without an NFS.