Blasphemy: I do not like the Tobin 104T Mods on my Bozaks

Note that I have Al tweeters, but paper mids - an uncommon mix. [...] Last week I put two bath towels over the mids - ah, much better. Today I rewired the tweeter L-pads to control the mid instead. Better. Much better.

I'm surprised that your B-209 need to be muffled. I doubt this is effects of the sonotube, which should be minor. Maybe resonance? Did you have the sonotue internally lined? I doubt that much matters, but the justification for the slanted sides of the flowerpot vs. the sonotube is to reduce standing waves and exciting resonance. No, I'm concluding this is not the problem.

Your midrange is not polarity reversed, correct? That will make it much too forward and altering the polarity makes it sound peculiar. (Yes, I've elsewhere posted Bob Betts' justification for why Bozak made that change, but for many of us a reversed midrange does not sound right.) Could this be the issue? See below.

Added an L-pad to the tweeters, hated them. Dammit. [...] Raised the front 1.5". Even better.

The paper-cone B-209 is 9 dB quieter than the (9 dB louder) aluminum-cone B-200Y or B-200Z tweeter. This causes imbalances, but not of the sort you describe. It would require attenuating the tweeter, not the midrange.

Do you still have the tweeters in front of the woofers? Dispersion might be the problem, but I'm thinking reversed midrange (see below) because you want it muffled a bit.

I'm going to run them this way for a bit and see. I still stand by my earlier post - I should have left them stock with a fresh re-cap. Well... maybe not. I will say that the Tobin mod did wonders with the woofer - much stronger and controlled sound - very impressed with that part of the mod. Moving the woofer/mid cross over was good too.

Tobin does three things: (1) moves crossover point down, (2) un-reverses midrange polarity change, (3) tape impedance rise of tweeter and attenuate for balance. Ok, that's sort of four things, but I'm lumping the tweeter mods together. The Zobel is not going to cause issues, but the attenuation does, which is why I have suggested switching to an L-Pad. Biggles had the same issues we all have had and reported the balance was much better with an L-Pad.

So the issue is, I think, either an un-reversed or reversed (it was proper polarity but was switched in the rebuild) midrange.
 
One more thing.

I suspect you can hear absolute polarity and you're hearing a driver polarity mismatch.

I suggest checking the polarity of the woofer, midrange, and tweeter to ensure they are all identical.

That would explain what you're hearing, particularly because the clarity of the Bozak will make it more apparent.
 
I have double checked polarity - they are all matching.

Tweeters are in stock location.

Maybe I really need to pull them apart and systematically double check everything for a 7th time. Something it uo with these.

All I know is that the mid is way too loud without the L-pad.
 
Hey man,

I feel your pain and am really sorry to hear about your sonic woes. I went through some real hell attempting to get my first Bozak 3-way crossover built. If it weren't for Retrovert's patience, I would have sold off the parts and moved on.

I read quite a few instance on the net, both old and new, where the OP states that they hate the Tobin Mod. But after a very short bit of sleuthing, it turns out not to be Pat Tobin's work at all. It's either a mistake in wiring (crossover or drivers), or the lack of a cover over the back of the midrange that's at fault. 98% of the time, it's the damned tweeters that are too dark. Which is an issue that would have been evident with a stock crossover as well. Rudy attenuated the B200Y tweeters down 9db as well. But having a blaring midrange being an issue, after checking and rechecking the wiring, hasn't happened to me in my travels. But I haven't spent any time with the paper cone B209, I can't speak from experience.

From our past discussions, I seem to remember that you double checked to make sure both your inductor taps are closest to the large inductor, correct? Yeah. Having the midrange tap on the lower lug will cause the midrange to be way, way too bright. It'll hurt. I know, I tried it. No bueno.

Other than what Retrovert has suggested, I got nuttin'.

Biggles
 
I have double checked polarity - they are all matching. Tweeters are in stock location.

Ok. That's removed some variables from the equation.

Maybe I really need to pull them apart and systematically double check everything for a 7th time. Something it uo with these.

I think you're good for one more disassembly.

All I know is that the mid is way too loud without the L-pad.

It makes no sense to require attenuation of the midrange, which is far lower power than either the tweeters (9 dB difference for aluminum vs. paper) or the woofer. Something is not right.

Check the inductor taps for the woofer. I suspect you have the woofer going too high and overlapping with the midrange, essentially doubling the sound at the low end. It would make sense that you've wired both the same way, hence overlapping in both speakers.
 
I went through some real hell attempting to get my first Bozak 3-way crossover built. If it weren't for Retrovert's patience, I would have sold off the parts and moved on.

Bozak Cult never leaves a Bozak owner behind.

It is worth noting that all of us have had issues when rebuilding Bozaks, and the knowledge base we have established helps to prevent others from spinning their wheels in the mud the way we have.

I read quite a few instance on the net, both old and new, where the OP states that they hate the Tobin Mod. But after a very short bit of sleuthing, it turns out not to be Pat Tobin's work at all. It's either a mistake in wiring (crossover or drivers), or the lack of a cover over the back of the midrange that's at fault.

That is the key. I wholeheartedly agree that the complaints about Bozaks have always turned out to be external to the speaker drivers and cabinet, and that the tweeter balance is an endless source of trouble.

What we do with Bozaks is essentially creating a new speaker from parts: new crossover, new wiring, adding a midrange enclosure, reversing the midrange polarity to match the other drivers, and moving the tweeters up and sometimes out of the box. This affords numerous opportunities for errors and omissions, as well as for unintended consequences like conductive flux residue, cold joints, and broken inductor wires.

It is major re-engineering, but the patient always lives if one is willing to do the necessary diagnostics and sometimes slap on the paddles a few times. I think we've reduced it to cookbook, and I have full confidence that this issue will be rapidly resolved, leaving another line in the knowledge base. (Which I must write up...)

98% of the time, it's the damned tweeters that are too dark. Which is an issue that would have been evident with a stock crossover as well. Rudy attenuated the B200Y tweeters down 9db as well.

Yeah, tweeter balance is a problem. So annoying. Sigh. There's also the tweeter placement. I do not favor low tweeters because my knees do not have ears and the dispersion is never optimal, neither is slot-loading a woofer the best solution.

But having a blaring midrange being an issue, after checking and rechecking the wiring, hasn't happened to me in my travels. But I haven't spent any time with the paper cone B209, I can't speak from experience.

I don't think this is a paper-cone issue. I think this is a crossover issue. Smells like it.

From our past discussions, I seem to remember that you double checked to make sure both your inductor taps are closest to the large inductor, correct? Yeah. Having the midrange tap on the lower lug will cause the midrange to be way, way too bright. It'll hurt. I know, I tried it. No bueno.

Ha! While you were writing that I was working on my reply and checking on dinner.

I think it's the other way around: the woofer is overlapping the midrange, reaching too high. The midrange is going down to 400 Hz, but if the woofer reaches up to 800 Hz, which it will do, that would add too much at the low end, particularly since a first-order crossover is a gentle rolloff.

It would be easy for both crossovers to be miswired so the woofer is running to 800 Hz instead of 400 Hz. That will provide an imbalance which matches the symptoms.
 
Agreed, from here it smells like inductor issues. Which makes a hell of a lot of sense. When I first began compiling information on building my first Bozak crossover, it was excruciating. I figured there would be one place where I could simply go and download the schematic and some notes. This wasn't the case at the time. All the information was spread out over the entire internet. I had no idea there was a 101T and the updated 104T. I bought everything to build the 101T at first. Once I had the correct 104T schematic, I attempted to find images that might help in the layout. The schematic gives the values for the inductors, but the Bozak N10102 crossovers do not. Which taps do I use? Well, oddly enough there are plenty of updated crossovers that show the midrange tap on the inductor for the older 101T schematic. This is no good for the 104T. This was only the beginning. There are so many ways this project can go wrong. The Bozak system is a logical fallacy for new owners who simply want to replace a few caps and get them into daily drivers.

Biggles
 
There are so many ways this project can go wrong. The Bozak system is a logical fallacy for new owners who simply want to replace a few caps and get them into daily drivers.

I know Biggles is aware of this, but others may not be.

If we look at the history this more or less makes sense. Bozak sold an upgradeable system, more or less, not speakers. The standard, as opposed to custom solutions or those created by third-party integrators, had single woofer, dual woofer, and the top of the line was quad woofers. One could purchase a variety of lower-cost two-way system (single or dual woofer with tweeters dragged into service as midranges) and add a midrange at a later point.

Being a small company, Bozak—this is an inference from its actions—did not want to maintain a multiplicity of drivers or crossovers. Since an 8 Ω quad could be built with four 8 Ω drivers (series to yield 16 Ω, then parallel to yield 8 Ω) this system could use the same 8 Ω crossover as the single driver one. The early dual woofer systems required putting two 8 Ω drivers in series to yield 16 Ω. (Later units used special 16 Ω woofers run in parallel.) But that meant a 16 Ω midrange was required, so from early on the midranges had both impedance. (The 16 Ω woofers are uncommon, as they were only used in the dual woofer configurations, or special 16 Ω quad configurations made by third-party integrators.)

So now two different crossover were needed: one for the 8 Ω systems and one for the 16 Ω systems. The multiplicity of early crossovers—N-101, N-102, N-103, N-104—were folded together into a single unit, the N-10102 (it is both an N-101 and an N-102, but it can also serve as an N-103 and N-104), which then used different taps on the inductors and then different taps on a capacitor bank. Very clever! A common set of parts could be sourced, a single crossover built, and all the magic would use taps. Later crossovers were hardwired and re-used the early numbers, but were essentially limited versions of the N-10102 which covered up the inductor taps with tape (I've unwrapped them to verify this) or used specific capacitor values without taps. This is why stickers were used to number those crossovers instead of silkscreening.

But then we come in fifty years after the development of the N-10102 swiss-army knife crossover and start tinkering with it. Oh, we say, we can use a 16 Ω tap to move the 8 Ω crossover point down. We can junk the capacitor bank in favor of fixed values, since we're building a specific speaker. And we can fix the polarity on a forward midrange. While we're at it, let's balance the tweeter with an attenuator and tame its impedance hump at resonance using a Zobel. And then there's a flowerpot midrange bell, which extends functionality in a way that only existed on the top of the line units.

So at this point we've built a new and more complex crossover using parts we had or ordered, and we've reconfigured the speaker as a whole. And that's where the trouble comes in.

Bozak gives us the freedom to do this upgrade, but if one doesn't have a background in loudspeakers and crossovers it can be difficult to get it all right. It isn't as cookbook as it would be with a kit from Heathkit, EICO, or Dynaco. Tobin was perhaps too close to the problem, in that he didn't really explain what he was doing. Perhaps because the people who were rebuilding the crossovers either were techs who understood the issue, perhaps because owners who didn't want to understand the math, they just wanted a cookbook solution, or perhaps because he was dealing with a small group of people and nobody ever asked for details. Either they did it themselves or paid him to do it.

I know that when I've explained the math behind the choices, and why one can't just drop a 16 Ω midrange into an 8 Ω system, I usually hear crickets chirping and frogs croaking.

So Bozak was never intended to be a modular system for non-integrators and those building new speakers from scratch, as we are. That's why so much discussion exists on the purpose for each of these modifications and their realization. This problem was created when people use Bozaks as a kit without realizing that there isn't much of an instruction manual. I suppose I should write one up incorporating the knowledge base and some diagrams.
 
OK, I have stripped the crossovers and I think I found the error. Someone correct me:

The woofer#2 wire connects to the top coil lug marked N-102/N-104 - correct?

The Mid #2 wire is on the top lug marked N-102 - correct?
 
It occurs to me, and is mentioned above, that the schematic is easy to find, but the directions are not, hence my post above to clarify.
 
Last edited:
Hey man,

I don't have the boards in front of me. But it sounds as though you're correct. Both inductor taps should be on the lugs closest to the larger inductor.

Oh wait, here's a picture. Notice the orange arrows.

Sexy1.jpg


Biggles
 
That is it, that is my error - mid are on the lower lug. I found a set of poorly done instructions that say for the paper mid to use the lower lug - clearly incorrect.
 
Note: For those who say they do not like the Tobin mods, what SiliconTI experienced is the factory setting on the inductor for the mid.

Awesome! Success rules, congratulations. And? You're welcome.

Biggles
 
Just returned and it looks like the dust has settled. It's a little late, so I'll only briefly respond.

At 8 Ω, the points are 400 Hz (full large inductor for woofer, N-102 & N-104 tap) and 2,500 Hz (half small inductor for midrange, N-101 tap)?

Given the crossover frequency for an inductor is calculated as:
f = R / (2 x π x L)
we can see that doubling L will halve the frequency, while halving L will double the frequency. So we want a small midrange value to move the midrange frequency higher and a large woofer value to move the woofer frequency lower.

If the full-value is used for the midrange, it will cross at 1,250 Hz (well, actually 1,157 Hz). So that will remove the upper end of the midrange, aside from what the tweeters provide via a first-order rolloff.

From what was originally described, it sounded like the woofer was doubling the midrange because the woofer was using the half tap instead of the full tap. That would also create comb artifacts as two drivers not within 1/4 wavelength were at the same frequency, but since the drivers are vertical that tends to not be as horrible sounding as horizontal drivers.

If the woofer was running with half the inductor it would cross at 800 Hz, and if the midrange were running at full the lower end would be 400 Hz (controlled by the capacitor) and the upper end be 1,250. So we'd have woofer(0 to 800 Hz) and midrange (400 Hz to 1,250 Hz). This will sound terrible. So I think this had to be the problem. Even if the midrange were correctly wired for 2,500 Hz, the doubling from 400 Hz to 800 Hz would be a problem, more so than might be expected because of the first-order rolloff.

It's rather late on the east cost, so the math may be off. (Special late-night math rules apply, even when integrals are not required.)
 
OK all, so, the woofer should be connected to the end of the inductor, (the lug marked N-101 on my crossovers) and the mid should be connected to the center tap of the inductor (the lug marked N-102 on my crossovers).

I have rewired the crossovers and will be installing them today. I have replaced the 7.5 Ohm resistor in the Zobel network with an 8 Ohm L-pad to better control the tweeters.

Film at 11.
 
Last edited:
Back
Top Bottom