Time-Alignment , Group Delay, and Distortion
Like any 8" driver, the B-800 does not well reproduce lower frequencies. So the issue becomes how much distortion accrues. The driver's position is likely within 1/4 wavelength, so the comb artifacts are not a significant problem.
The time-alignment — which varies with frequency — between the B-199 and B-800 will be different because the acoustical center for the two drivers is different.
How well can humans hear that? Well, I'd point to the Dahlquist DQ-10 and other time-aligned drivers.
Linear power distribution requires that the area of the radiating surface match the wavelength lengths being reproduced. This is why woofers cannot reproduce high frequencies, and small drivers cannot properly reproduce low frequencies.
Bass-Boost Unnecessary
For one thing, you concluded the additional boost is probably not needed.
Correct.
First, the B-199 clearly requires no bass boost; that is gilding the lily.
Second, the 8" B-800's output is going to be far smaller than a 12" woofer, just look at the cone area: Area is proportional to the square of the diameter, to 64 vs 144 (note, the actually area has
π in it, but since that is constant for the two we can neglect that) a factor of 2.4 difference in the area.
Impedance Shift from Parallel Drivers
But also, you refer to the out-of-band signal passing through the driver possibly causing other issues beyond the impedance drop. Questions about that:
A. Does the "other issues" refer to the B-800A cone breakup you discussed earlier in the below 100Hz band? And/or possibly even occurring above 100Hz? Or something else?
B. How does impedance drop come into the picture? Is that an artifact of the cone being immobilized below 100Hz? Apparently though a reduction of back-EMF? That's what I'm reading into it, that the impedance drop occurs at or below 100Hz?
Not quite. Back-EMF has no bearing on this situation. That does not alter the overall speaker's (aggregate) impedance.
Back-EMF arises because, as Faraday observed, moving a wire coil in a magnetic field induces current flow through that wire. Mechanical energy converted to electrical energy. A motor in reverse. Basis for alternator and generators. So as the cone returns to the rest position a current through the voice coil is generated, aka back-EMF. This back-EMF is absorbed by the amplifier. Smaller drivers have smaller voice coils and smaller excursion, thus generating smaller amounts of current. The core issue here has nothing to do with the surround damping out the lower frequencies and cone movement. Think, instead, about current flow through the voice coil, because that's all that matters from the standpoint of impedance.
Without any filter for the midrange the current flowing between the amplifier terminals, i.e. flowing from one side to the other (it's AC so we can't say from – to +, only from terminal to terminal) therefore simultaneously passes through the voice coils in
both drivers at the frequencies where the drivers overlap. The drivers, therefore, are in
parallel, which is now half the impedance for the overlapped region, which will shift the loadline for a tube amplifier, but does not matter for solid state, provided, of course, that the impedance does not too low drop.
Remember, the B-199 has an inductor so current only flows through its voice coil, again subject to rolloff, below 400 Hz. The B-800A used as a midrange has no high-pass filter, so current flows all the way down to 100 Hz and below that down to DC. The cone's lack of movement does not stop current flow through the voice coil.
The crossover electrically isolates the different drivers which is why the impedance remains constant even though three drivers are present. Subject to overlap from rolloff, of course, but as one driver is rising the other is falling so the impedance curve remains flat. (Yeah, not exactly, because driver impedance is a convenient fiction, but you get the idea.)
Midrange Enclosure Now Necessary
Also, I believe in order to add bass boost with the midrange driver, it probably needs not to be under a salad bowl in order to be working into adequate volume. But to provide the best sound, isolated from the B-199A, it needs to be under the salad bowl. Something of a paradox. That would seem to argue against overlap.
Yes, the cover is too small for linear bass response, i.e. below the normal midrange band which would occur using the high-pass filter, and the backwave would likely cause some distortion.
The question, again, is what is the driver rolloff between 100 Hz and 400 Hz? If the B-800 tends to not properly reproduce those frequencies the overlap is small and the B-199 dominates. Since the Concert Grand has four B-199s, it stands to reason that the contribution of the B-800, and any ensuing distortion, is small. But small enough to be invisible? Probably not if one had the A/B comparison. The B-800 system may simply have better dispersion of the midrange by way of the larger cone, so the perception may be that of a louder midrange, which is where humans best hear, and this increase in volume may be perceived as desirable, much like an inverted bathtub curve on an equalizer. Look at the purported benefits of the inverted polarity midrange for those who cannot hear absolute polarity.
Removing the cover would cause the very cross-modulation, i.e. between woofer and midrange, we now prevent with the flower-pot cover. That cover, again, was not put into the smaller Bozak cabinets to reduce manufacturing costs. Or so I would assume. I can see not benefit to the sound for omitting it, and we see the benefits of using the flower pot. Seems to be purely an economic issue.
So the B-800 would require an internal cabinet suitable for 100 Hz, basically the same type of cabinet one would use for a B-800 operated as a full-range driver as in the Sonata 1 or the B-801. The small B-800 models — LS-200 and DMS 2500 — are both ported cabinets. The ported versions produce less, but more accurate, bass with the port plugged, but the 1 to 2 cu ft cabinets are a bit too small for the driver used as a woofer. Any substantial cabinet will rob volume from the B-199. Less critical in the large Concert Grand, more critical in the smaller B-302.
Speculation that Economics, Not Sonics, Was the Reason for Elimination of Midrange's High-Pass Filter
The whole concept of the capacitor-less filter for the midrange may, in fact, be based purely upon the high expense of PIO capacitors of 1950s and 1960s, not merely the low performance which was not particularly good, even when new. Our polypropylene capacitors are light-years beyond the 1960s PIO.
I expect the goal was eliminating the expense of a large bank of capacitors necessary for the midrange. The cost for a 50 µF capacitor in those days would have been high. In 1963, Lafayette lists an oil-filled metalized capacitor of 8 µF at $1.19, so figure $6 for the full bank, but Bozak would have paid about 1/3 of that wholesale, so about $2. That's still expensive: about $40 in today's income. (The B-800 cost $50 at the same time, so the capacitor would have been 5% of the cost of adding the midrange, plus, of course, the expense of the inductor.)
Remember, inflation has been terrible. In 1960 the median family income was $5,620 (I looked it up) or $107 per week, and that was pre-taxes. So figure that median family kept about $75 after federal and state. Housing was less than one third of what it was, in constant dollars. For purposes of ready comparison, a skilled factory workers made somewhere around $2 per hour (pre-tax) and the minimum wave was $1.25 per hour (pre-tax).
A non-polar electrolytic would have been less expensive, of course, but would have had a terrible sound. So that was the Hobson's Choice which Bozak was attempting to circumvent.
TL/DR
Retovert, it looks like you were earlier thinking that woofer/midrange overlap could be a good thing, but later apparently changed your mind. At first, you seem in favor of the idea:
Mmmm, not so much. I never was in favor of it, just observing that it might not be entirely as bad as I initially thought.
I wanted to clarify if you think the overlap is a bad idea:
A. Mainly for the B-800;
B. Mainly for the B-800A;
C. A bad idea period.
Yes, Yes, and Yes! Three chances to win, and you hit all three! Woo-hoo!
I think this is enough to kill the idea that running the midrange as a full-range driver to eliminate the capacitor is a poor choice, both for the economy of eliminating the high-pass filter (capacitor) and foolhardy from the perspective of sonics. I think this was done purely to save money, not because of the sonics. Remember, a modified B-209 could have been generated if this was such a great idea.
You could, of course, try the B-800 both with and without a polypropylene capacitor, use a switch for proper A/B testing, and see what happens. I'd be surprised if the no-capacitor version sounded better.
