Thursday, May 28, 2015

Bozak Rebuild Project - Part 11

It's been over three years (Yikes! How time flies) since I started this quest to rejuvenate my 1958 Bozak 302-A speakers and slowly but surely things have gotten better. New satellite enclosures, new network designs, new drivers, new inductors, and new capacitors have all improved the original sound that today rivals the best of the best high-end designs. Along the way I have shared what I discovered with you in hopes that your own learning curve for improving your own speakers will be shortened.

This segment of my ongoing saga recaps the current thoughts on the redesign of this classic loudspeaker. I will share with you the driver complement, constraints, and challenges in maintaining a high 1-watt sensitivity loudspeaker while still maintaining excellent accuracy in both the frequency and phase domains.

First, I wanted to share with you why I chose the Bozaks over other models: their high sensitivity. This translates into loud music with modest amplifier power. The Bozaks have a 1-Watt sensitivity of about 97dB/W/m meaning that it takes A LOT less power to drive them to the same SPL as an 87dB/W/m loudspeaker. (I’ll talk about the pros and cons of low and high efficiency loudspeakers in another segment since that’s another can of worms.)

I started with the philosophy of making as few changes to Rudy Bozak’s original 1958 design as possible but rather applying current technology to a dated approach. I initially tried the well-regarded “Pat Tobin Mod” of upgrading capacitors with positive results. However, measuring each individual driver with its original first-order network revealed each driver severely “bleeding” beyond its desired operating band. So higher-order crossover networks seemed to be the solution.

Second and third order networks were tried with mixed results and finally I settled on fourth-order Bessel networks for the M-T and T-ST but a second-order Butterworth from the W-M. Along the way I discovered that my Bozak tweeters had a lot of holes in their diaphragms and these tweeters gently rolled off at about 8KHz. So replacement tweeters with the same high sensitivity were sought out. This proved to be a long process since most drivers in this range were PA speakers that were not what one would consider high-end.

I finally came across the Peavy Versarray that used a planar tweeter with very good sensitivity, the RD-1.6. After many attempts at taming this driver, I abandoned the work. The next choice was the Mundorf AMT 164UM2.1 that is still in the system today. It performed much better than the Peavy however it still had issues with the top octave. Using a forth-order Bessel LPF to limit the HF content turned out to be a good choice but that meant searching for a super tweeter.

Again the search mainly found PA drivers and exotic compression horns that – while they may have sounded good – were beyond the reach of my wallet. Using Panasonic EAS-10TH400A leaf tweeters in a previous loudspeaker with fantastic results I decided to search for something using leaf technology.

I purchased a pair of Bespoke Aria tweeters (these folks use the Monacor RBT-95 true leaf super tweeter) but I never liked the way they sounded. This has been my main efforts over the entire year: taming the Monacor. I had faith that all I needed to do was to persist and persevere and things would work out. But until I made some near-field measurements, I did not completely understand why the sound was so weird. Anyway without belaboring the point, the solution was to cross this super tweeter REALLY high (currently it crosses at about 34KHz) to control the Monacor's contributions below the crossover frequency that really screwed up the sound of the stellar Mundorf AMT tweeter.

Monacor RBT-95 Leaf Diaphragm

Below is a photo of the final crossover network (as of today) and here is the configuration:
  • Super-Tweeter high-pass filter (HPF) uses Russian Teflon capacitors shunted with Corning glass capacitors with wire-wound ground-path (shunt) chokes
  • Tweeter uses a combination of Mundorf, Audyn, Russian Teflon, Obbligato, and Corning glass capacitors with ribbon signal-path chokes in the low-pass filter (LPF) and wire-wound chokes in the HPF. A sort of "Zobel network" is also used to minimize T-ST interactions.
  • The midrange uses Mundorf, Russian Teflon, Clarity, and Corning glass capacitors with a Solen capacitor for the notch filter. The midrange network uses ribbon chokes in the signal path and wire-wound shunt chokes. The midrange uses a notch filter at about 3KHz to help smooth a mechanical resonance. A Zobel network is also used.
  • The woofer (actually two woofers in an isobaric arrangement to improve bass response in a small cabinet) uses a low-loss toroid inductor and Clarity caps for the crossover and the Zobel network uses Clarity and Audyn caps.
  • All signal-path resistors are Mills non-inductive.
  • The midrange and tweeter are T-pad attenuated. All attenuators are better than 1% hand-matched.
  • All capacitor and inductor values in the network are spot-on hand-measured.
  • All impedances of drivers at the crossover points were confirmed via LRC resonant measurements with my HP 465 oscilloscope and function generator.
  • All network LPF/HPF performance was verified with near-field measurements of each individual driver and then its partner to assure minimal driver interaction.
  • All point-to-point network wiring is #14AWG stranded silver-plated OFHC copper, Teflon insulation.
  • W-M crossover is -6dB BW2; all others are -6dB Bessel4 (hence the high parts count for this 4-way system).
  • Each driver has its own network board and speaker terminal pair.
  • The satellite box is now a truncated trapezoidal-pyramid design with curved front edges and adhesive felt on the face
  • Presently each speaker is tri-wired (one amplifier, two channels, six speaker wires total). Wiring varies in material content but all are star-quad.

Crossover Network as of 5-23-15

Remember that in my world things routinely come and go but I can report that I am pretty happy with the way they sound, and that is saying a lot. To give you an idea of what they sound like, give a listen to the outstanding Vapor Sound Joule. The Joule has much better bass control and my Bozaks have a better top octave. The lower mid-bass between the W-M is the other noticeable difference but this difference depends a lot on room resonances and speaker placement.

There is one other change that I am in the process of implementing: better crossover-to-speaker wire terminations. It is well known that ferrous terminations (steel nuts and bolts) will magnetize slightly as the signal passes through. Magnetizing the nuts and bolts steals some of the energy that really should go to the speakers. Using stainless steel versions changes all of that since stainless steel has no appreciable magnetic properties (see below). As you can also see, I continue to use star-quad stranded silver-plated-copper wiring throughout making the biggest difference in the size of the sound stage and other differences in transient response.

Stainless Steel Nuts and Bolts

Also evident in the above image are the numerous holes from countless failed attempts at a HPF design, redesign, re-redesign, re-re-redesign...for the super tweeter. One day I will replace that board and make it at least look better. And some other day I will design and build a better woofer enclosure (it's already on the drawing board waiting for spousal approval).

BTW, many of the ideas I used in this redesign you can also use on any design. Here is a brief summary of generic mods:
  • Take out your crossover network and go through it by using better wire in a point-to-point method abandoning those puny copper PC board traces.
  • Use good wire everywhere and orient the chokes 90 degrees to each other (horizontally, vertically, or laterally) and don’t place two similarly oriented chokes near each other.
  • Use good quality capacitors in the crossover network
  • Use three screws (wherever possible) in a triangular pattern to mount the drivers. BTW, you can test this on your woofers by loosening all but three screws are still in the holes but not touching the basket. Give this one a shot and be amazed.
  • Put felt around the midrange and tweeter drivers and make sure that the screws used on them are as flat with the baffle board as possible
  • Use separate ground runs from the back speaker terminals through the crossover network and then to the driver
  • Try bi-wiring or even tri-wiring and experiment with wire designs of your own. I prefer star-quad with silver wire; you may have a different preference since this is a pretty lively wire and may cause your mid/tweet to sound harsh or metallic.
  • Put your speakers as far into the room as aesthetically possible without breaking up your relationship (1/3 of the way toward the listening chair is a good number)
  • Move your speakers around a little here, a little there, and see if the bass or sound stage improves
  • When finished moving them around, put your speakers on spikes (three if possible; two in front and one in back). If you have hardwood floors, slip three pennies underneath (again two in front, one in back) and be amazed! I call this “Phil R’s patented 6-cent fix.” Catchy, right?
  • Buy a really good power conditioner and plug all of your equipment into it (no sharing of outlets). Isolation transformers are not enough; RF filtering is absolutely essential.

I’m sure there is more but this should keep you busy for a while (maybe three years or so ;). Tweaking is part of the fun of this hobby and every point of the playback chain has the potential to be tweaked. Remember that swapping out speakers will give you the biggest change in your system. If you are not happy with them, start there.

So how does it measure? I knew you were waiting for this. Well this was a surprise to me: it measured - from the same listening position I have made all of my other measurements - rather disappointing since the microphone I used suffers greatly above 6KHz and below 300Hz. Many of the 3,000 or so measurements I made over these three years looked great on paper but sounded absolutely awful. This one shown below measures fair but it sounds truly amazing. To give you an idea of how the sound changed, the oldest surviving RTA measurement (hard drive crash at fault here for eliminating older measurements) using the original four-tweeter design and the Pat Tobin network is shown for comparison. Also shown in green on the graphs are the approximate crossover points for each of the two systems.

The characteristic sound of the original Bozak is gone except for the captivating midrange of the B-209B. The woofer box is now ported so the bass is quite a bit deeper; it also has a slight “tubed” port characteristic. While the flared port is not objectionable, it is another area of interest to me. A tapered port design may be in the distant future. For now, I am pretty happy with the way it sounds. With the exciting dynamics and deep bass my wife is also. So stay tuned and see what pops up out of my head next in Part 12?

See also Part 1, Part 2, Part 3, Part 4, Part 5, Part 6, Part 7, Part 8, Part 9, and Part 10

Yours for higher fidelity,

Philip Rastocny

I do not use ads in this blog to help support my efforts. If you like what you are reading, please remember to reciprocate by purchasing one of my eBooks or through a PayPal donation, My newest title is called Where, oh Where did the Star of Bethlehem Go? It’s an astronomer’s look at what this celestial object may have been, who the "Wise Men" were, and where they came from. Written in an investigative journalism style (like that of the Discovery Channel), it targets one star that has never been considered before and builds a solid case for its candidacy.

My other titles include:

Copyright © 2015 by Philip Rastocny. All rights reserved.

No comments:

Post a Comment

To comment on this blog, you must first be a member. All comments are moderated.

Note: Only a member of this blog may post a comment.