Looking at the crossover, Dennis added a 0.5 mH low pass inductor. Looking up the Jantzen used, it adds a dc resistance of 0.5 ohms right in series with the woofer. From
@amirm measures of the originals, the original system impedance was ~ 5 ohms below the BDC frequency range and above port tuning frequencies. So, this added inductor will add ~ 0.8 dB loss to the ~ 200 to 400 Hz range. The rest of the difference in the bass will be due to
- Dennis added sonic barrier which will somewhat lessen port coupling efficiency, reducing sensitivity below 200Hz. This probably is the main contributor to getting rid of the 70Hz peak as well
View attachment 65662
- if not from the same production batch, raw drivers could easily have unit to unit variation of a dB, making up the rest
The vertical off axis dip in the response curve gives away the crossover frequency as ~ 2.7 kHz (eyeballing it from the graph):
View attachment 65665
Its also a fairly steep notch, also giving away a fairly steep high pass.
The crossover schematic confirms this, being 3rd order electrical (18 dB/octave asymptotic). The 2.2uF cap helps get rid of the tweeter impedance peak at resonance, keeping the crossover rolling off:
View attachment 65666
Checking out the hificompass test shows that the Vifa tweeter's distortion @ 4V takes off below ~ 2.5 kHz:
View attachment 65663
4V would be ~ 96 dB sensitivity/2.83V at one meter. From the data sheet:
View attachment 65664
(above 2 khz, the tweeter operates in half space on the baffle)
So, putting this all together, the third order crossover at 2.5kHz to 3 kHz is a very sensible choice to keep the tweeter distortion in check all the way to quite high playback levels.
I think Dennis made another really good choice elevating the on axis slightly from 800Hz to 3 kHz, to help compensate for the woofer's off axis roll in this range (as shown by the DI curve). I find this definitely helps make the spectral balance a bit more neutral in my designs. This is where allot of Dennis' voicing black magic comes in, because there are no hard and fast rules how much extra on axis response is needed to compensate for the roll off off-axis (and it can depend on room placement a bit), but IME and looking at the dispersion, I would expect this to be about perfect.
View attachment 65669
I also think its great that Denis was able to keep cost down by not replacing any big inductors, just adding one small lower cost one. This choice would have reduced the amount of design freedom he had available, so it makes this result even more impressive.
Looks like a collection of great design choices, but no surprise there.