forums.stevehoffman.tv/threads/tobin-mods-to-bozak-302a-urban-speakers-definitely-worth-it.75950/page-2
The early Bozaks ('50s) were never an easy first sale. I worked as the repair tech in a prominent stereo store in the '50s (yes, I was barely out of diapers) and saw this first hand: In A-B comparisons in the listening room, most first-time buyers would opt for a competing system that had more flash bang in the sound. The Bozaks are just too true, while the competing iron jumped at you. Bozak was under a lot of pressure from the dealer network to make the Bozaks sound more like the inferior competition which was selling better. Most first-time buyers, after learning how little they liked hearing the same flash and bang in everything they played, would come back to the store in a few months and say, "Uh...could I hear those Bozaks again?" But this was little consolation to the dealers, who wanted Bozak to compete for first-time sales.
The early Bozak systems were perfectly balanced - woofer to midrange to tweeters - and everything in phase. So in the early '60s, when the aluminum-coned B-209A/B midrange and B-200Y tweets came on stream, Bozak really fouled up the xovers. For starters, the polarity of the midrange was reversed. This resulted in a mid sound that was incredibly prominent without actually being louder. With the mid reversed in polarity compared with the woofs and tweets, the xover regions were cancellation notches. This 'isolated' the midrange in a manner similar to a picture which is surrounded by a wide, plain matte - it makes the subject more prominent. But the down sides were degraded overall smoothness and seriously degraded stereo imaging. Without any doubt, the reversed polarity on all Bozak N-10102A xovers should be corrected. (It is not crossed on the N-10102 xover, used with the early paper-coned B-209 midrange and B-200X tweets.
The other issue was the Y tweets. They are about 9 dB hotter (louder) than the woofer and midrange (and the previous paper-coned X tweets). Presumably this was to enable inclusion of a Brightness control which could reduce tweeter level or make it higher than flat. Only a very few systems actually had the Brightness control, mainly the early Symphonies and a few B-302A systems. After just a year or so, the Brightness control pots were eliminated. A simple network consisting of a 25 ohm resistor paralleled by a 2.0 uF capacitor was put in series with the Y tweets. The results, by today's standards, are pretty awful. It results in a big hump in the mid highs, 5.0 kHz to 10.0 klHz, allowing the natural rolloff of the Y tweet above 10 kHz to go unaided. Can you say, "Disco?"
It is not my job to help Bozak sell speakers in a tough '60s and '70s market. What I am doing is re-engineering the xovers to remove the strange tweaks and allow the world-class drivers to sing in their full, true voices, unhampered by the craziness of the '60s and '70s. To that end I designed a much better circuit to drive the Y tweeters. It is more complex than the bozak one, and works much better. It reduces the extra 9 dB level to match that of the woofs and midrange, flat, without the hump. Then, with 9 dB of 'extra' level to play with, it is used to extend the extreme high range, boosting from 10 kHz up to the tweeter's normal limit of about 16 kHz. The result is highs that are very smooth; no peaks, no dips, and a smoothly extended upper range.