Historically speaking in early 2-way systems, woofers were permitted to roll off under the influence of the voice coil inductance and tweeters were used with simple first-order networks (BW-1). This kept efficiency high and phase distortion to a minimum and this approach is still used today in budget loudspeakers. But other highly undesirable effects rear their nasty little heads in too simple of a design and engineers initially opted for a passive network component on each driver to better limit the bandwidth of the driver.
The more sophisticated the network (second-order, third-order, etc.), the more power is "stolen" by the network itself (insertion loss) and the speaker's 1-Watt sensitivity is therefore reduced. But the benefits outweighed the loss in efficiency and so began the evolution of "modern" crossover network design (I use this term loosely since there is great debate between what is and what is not the best crossover network design to use). Today, 2nd, 3rd, and 4th order networks are routinely used in well-designed speakers but each design corrects something while compromising something else. In other words, there is no such thing as a free lunch.
Rudy Bozak did not like the sound of higher order crossover networks and as a result all of his systems used BW-1 designs. While such designs can be well implemented (and his was done pretty well considering its era where esoteric capacitor designs were in their infancy), knowledge about loudspeaker behavior is much better understood today than in Rudy's time. So with this understanding and with a wealth of new technology at my disposal, I decided to see what could and could not be done to improve this early design.
Others have upgraded the existing network, most notably Pat Tobin's work with his B-101T design. Here Pat attenuated the over-anxious B-200Y tweeter by -9dB and then bypassed the attenuator at the proper point (the 2.0uF capacitor) to recover the extreme HF content (above 8KHz). He also used the existing chokes although changed the value of the midrange component (selected the other available tap). Lastly, he lowered the midrange crossover point and as a result the speakers sounded much like the original but much more balanced. Below is the schematic for this modification to a stock N-10102 network.
Applying this modification to my own Bozak crossover network (mine is a slightly different version of a standard B-302A) yielded the following RTA measurement. A Zobel network is used on the woofer and a 2nd order network on the midrange LPF. (it measured a significant contribution above the designed crossover point with the 1st order network). Much like Mr. Tobin, I decided to try to use the existing inductors where possible. Note that instead of the usual two-tweeter arrangement, three tweeters were used changing the acoustic output in that band and therefore further attenuation was required. Also added was a VIFA BC25SC55-08 super tweeter to attempt to recover the top octave. The problem with matching an audiophile grade super tweeter to the Bozak system is that the Bozaks are highly efficient (98dB/W/m or more for the bigger Bozak designs). Even with this super tweeter the top octave still sounded quite weak compared to the rest of the spectrum.
Note that while taking the RTA measurement, the Peak dBFS scale somehow messed up (estimated to be about -20dBFS).
While interesting to listen to and very pleasant while retaining much of the original character of Rudy's design, they still sounded mediocre. The transients were superb but there was just too much subjectively perceived IM distortion clouding the sound stage and making it quite two dimensional. Even changing polarities and hand-matching components did not help to make it that big sound I suspected it could deliver. I believed that this network just did not permit the drivers to perform to their full potential. The original idea of using the existing inductors had to go in order to change the crossover points and to move to a 2nd order network..
For the past 18 months, I have redesigned, tested, retested, removed, replaced, repaired, reworked, and rethought about every aspect of a second-order Bessel design I could for these drivers. Giving up on the VIFA super tweeter, I turned my gaze to the professional market in hopes of a reasonable marriage between efficiency and sensitivity. Two drivers caught my attention, the Peavey RD1.6 and the Fountek Neo Pro 5i. Both drivers showed promise and I settled on the Peavey since musicians themselves reviewed that it sounded more like a home speaker than a PA speaker (I thought this was a good thing).
The latest design uses a 2nd order Bessel on the midrange LPF and HPF, a Zobel on the woofer and midrange, and a 2nd order Bessel on the Peavey super tweeter. This is NOT a finalized design; I plan to also swap out the tweeter network to a 2nd order LPF and HPF Bessel and see what can be done to encourage more from the B-200Y's 8KHz wall. Below is the last RTA of the system.
Now the sound stage is H-U-G-E, deep, wide, and tall. As you can see, the top octave is back but the Bozak B-200Y tweeter is still dying near the 7KHz crossover point. Removing the tweeter LPF causes zero difference in the RTA measurement (sigh).
This is a network that has a nice wife acceptance factor (WAF) where before she thought it sounded good but now she thinks it's great! Ambiance and room resonances in live recordings are much more revealed with the return of the top octave and the 2nd order Bessel seems to have tightened up the midrange driver's sound. Vocals are more distinct such as Enya's Watermark album was blurred now background voices are as clear as hers. Details in this album's noise floor are readily revealed and nuances in piano soundboards are also more evident.
Putting a 1.5uF kicker capacitor on the tweeter helps the RTA graph look better (much like the 2.0uF capacitor used by Tobin) but ruins the sound raising vocal and instrumental textures to a near tizzy state.
So the next round of mods will address the tweeter. Foil chokes are on order and should arrive in house in a week or so. I plan also to change the woofer LPF also to a 2nd order Bessel in the near future and that should round out the network mods. So stay tuned and see what pops up out of my head next in Part 10.
See also Part 1, Part 2, Part 3, Part 4, Part 5, Part 6, Part 7, and Part 8
Yours for higher fidelity,