augerpro said:
Hmm ... looking at the factory curves of the B&C 6MDN44, the twin peaks at 4.5 and 5.5 kHz look ominous. The peaks re-appear in the impedance curve, indicating they are very real and seriously high-Q in nature. I also see an impedance ripple at 1.2 kHz, which indicates something is going on there as well. It takes a lot of resonance to punch through the heavy inductance of an impedance curve.
The esteemed Zaph is OK with multiple notch filters, but they give me the willies. Too complex, too much fine-tuning for individual drivers, too many parts that cost a lot of money for decent quality - all fine and dandy for digital-EQ crowd, the 24 dB/octave active-filter users, or "ignore it and it will go away" folks, but troubling for me. I'm very serious about drivers that don't require prosound filtering or EQ for low-coloration response.
The 18Sound 6NMB420 does look interesting. Pretty decent Xmax for a mid/midbass driver, and the very low inductance of 0.1 mH is a significant point in its favor. Still trying to figure out the exact difference between this driver and the 6ND410 midrange - Mms is almost the same, Xmax is a bit less, inductance is much more at 0.67 mH.
My guess: the 6ND410 voice coil is shorter and the "single demodulating ring" is removed, thus tightening up the gap, raising the flux field, and delivering the 2.6 dB gain in efficiency. The impedance curves for the two drivers also look odd - it looks very much like the impedance curves have been inadvertently swapped - the 6NMB410 looks noticeably less inductive (contradicting the printed Le specification) thus raising a question about the FR curves possibly being swapped as well.
Attachments
Lynn,
I just looked at that site
http://www.picosound.de/D_soft.htm#tspchk
and when I calculated right, the double Mms of the Audiotechnology (Skaaning) drivers compared to typical pro drivers only means -1.5dB in sensitivity, what could easily be compensated by choosing Fs and Qts exacly as needed.
I just looked at that site
http://www.picosound.de/D_soft.htm#tspchk
and when I calculated right, the double Mms of the Audiotechnology (Skaaning) drivers compared to typical pro drivers only means -1.5dB in sensitivity, what could easily be compensated by choosing Fs and Qts exacly as needed.
Hmm. The ripple in the impedance curve of the purported 6NMB620 does correspond to the main breakup peak in the FR curve, so at least impedance and FR curve seems to belong together. Also if you look at impedances at 20k, the purported 6ND610 does have a lower impedance than the purported 6NMB620. But who knows...
Detective Work
When I worked at Tektronix as a technical writer, graphics mixups were more common than you might think for a 20,000-employee corporation that made a wide range of MIL-SPEC instruments. As the writer responsible for the final manual that shipped with the product, you really had to bird-dog the production people at every stage, otherwise about 1 graphic in 20 would be swapped, causing no end of confusion on the user's part and many costly phone calls and visits by the field office people to remote customer sites. Very embarrassing for the writing team to print the inevitable "change pages" that were inserted into the printed manual.
Even today, with Web and desktop-graphic production, graphic mixups are more common than you might think - you even see it in 4-color mass-circulation scientific magazines. Of course, I'm on the lookout for it, having done my share of writing, editing, and production work. The momentary faltering of the editorial hand is immediately visible to anyone that's done this kind of work. (Yes, I notice typos, and it takes a little self-control to not mark up a book with a blue pencil.)
Reading the specs of the 6ND410 midrange and 6NMB620 midbass closely, it is evident they are closely related - the similar Mms spec gives that away. With the Mms so similar, the difference could come down to little more than a different VC former and a different dust cap (visible from the pictures).
It is unlikely the text describing the "single demodulating ring" and the Le of 0.1mH of the 6NMB420 are both wrong. This kind of thing is usually at least glanced at by the chief engineer and the company president before it goes public, so the chances are good the text is accurate.
We also see no mention of a "demodulating ring" in the 6ND410, a Le of 0.67 mH, and a 2.8 dB higher reference (Theile/Small) efficiency, which is consistent with the pole-pieces coming closer to the VC, thus increasing the flux in the gap. Since the Mms and emissive area are pretty much the same, the only difference in efficiency can come from increased flux in the gap - the result of a shorter voice coil and the absence of a conductive shorting ring.
A driver either has more inductance or it doesn't. The curves are a little hard to read because the real inductance is more complicated than a simple L + R. For precise correction of inductance to 100 kHz or higher, it take a two sets of Zobel filters, not one, and even then the approximation isn't quite exact. (Back in my Audionics days, I impedance-compensated some of our loudspeakers out to 500 kHz, to make a kinder load for unstable transistor amplifiers. This required two sets of Zobel compensators for the tweeter.)
The breakpoint for the published 6ND410 curves is visibly higher than the breakpoint for the 6NMB420, as well as the 20 kHz figure being different, as MBK has astutely noticed. As he also mentioned, the FR curves are correlated as well, with the published 6NMB420 impedance curves showing ripples at the exact same frequencies as the sharp peak at 5.5 kHz in the FR curves. Thus, I think it likely both the FR and impedance curves for the 6ND410 and 6NMB420 were accidently swapped in pre-production.
It also interesting to look closely at the 15MB700 and 15NMB420 twins. Both have 3" diameter voice coils and more significantly, identical Mms of 73 grams and Fs of 42 Hz. This points to the same cone, and probably the same surround, spider, and VC former.
The curves for the two 15" drivers are very similar but - not quite identical. This comes down to differences in the reflections from the basket, magnet structure, VC venting pipe-mode resonances, and very likely, measurements taken at different times with slightly different test protocols.
Yes, fine points, but correlating specs with the physical driver is very important. The specs at best are a very indirect pointer to the complexity of the real-world system. The real values of specs is telling us something about how the driver is made - this, by the way, is why I like to see drivers tested on large flat baffles, so I can see the driver, not the baffle artifacts. The "differences" between the 15MB700 and 15NMB420 could be nothing more than different-sized baffles - indeed, baffle diffraction alone could account for the ripples I complained about in the Supravox CSD measurement.
MBK said:
When I worked at Tektronix as a technical writer, graphics mixups were more common than you might think for a 20,000-employee corporation that made a wide range of MIL-SPEC instruments. As the writer responsible for the final manual that shipped with the product, you really had to bird-dog the production people at every stage, otherwise about 1 graphic in 20 would be swapped, causing no end of confusion on the user's part and many costly phone calls and visits by the field office people to remote customer sites. Very embarrassing for the writing team to print the inevitable "change pages" that were inserted into the printed manual.
Even today, with Web and desktop-graphic production, graphic mixups are more common than you might think - you even see it in 4-color mass-circulation scientific magazines. Of course, I'm on the lookout for it, having done my share of writing, editing, and production work. The momentary faltering of the editorial hand is immediately visible to anyone that's done this kind of work. (Yes, I notice typos, and it takes a little self-control to not mark up a book with a blue pencil.)
Reading the specs of the 6ND410 midrange and 6NMB620 midbass closely, it is evident they are closely related - the similar Mms spec gives that away. With the Mms so similar, the difference could come down to little more than a different VC former and a different dust cap (visible from the pictures).
It is unlikely the text describing the "single demodulating ring" and the Le of 0.1mH of the 6NMB420 are both wrong. This kind of thing is usually at least glanced at by the chief engineer and the company president before it goes public, so the chances are good the text is accurate.
We also see no mention of a "demodulating ring" in the 6ND410, a Le of 0.67 mH, and a 2.8 dB higher reference (Theile/Small) efficiency, which is consistent with the pole-pieces coming closer to the VC, thus increasing the flux in the gap. Since the Mms and emissive area are pretty much the same, the only difference in efficiency can come from increased flux in the gap - the result of a shorter voice coil and the absence of a conductive shorting ring.
A driver either has more inductance or it doesn't. The curves are a little hard to read because the real inductance is more complicated than a simple L + R. For precise correction of inductance to 100 kHz or higher, it take a two sets of Zobel filters, not one, and even then the approximation isn't quite exact. (Back in my Audionics days, I impedance-compensated some of our loudspeakers out to 500 kHz, to make a kinder load for unstable transistor amplifiers. This required two sets of Zobel compensators for the tweeter.)
The breakpoint for the published 6ND410 curves is visibly higher than the breakpoint for the 6NMB420, as well as the 20 kHz figure being different, as MBK has astutely noticed. As he also mentioned, the FR curves are correlated as well, with the published 6NMB420 impedance curves showing ripples at the exact same frequencies as the sharp peak at 5.5 kHz in the FR curves. Thus, I think it likely both the FR and impedance curves for the 6ND410 and 6NMB420 were accidently swapped in pre-production.
It also interesting to look closely at the 15MB700 and 15NMB420 twins. Both have 3" diameter voice coils and more significantly, identical Mms of 73 grams and Fs of 42 Hz. This points to the same cone, and probably the same surround, spider, and VC former.
The curves for the two 15" drivers are very similar but - not quite identical. This comes down to differences in the reflections from the basket, magnet structure, VC venting pipe-mode resonances, and very likely, measurements taken at different times with slightly different test protocols.
Yes, fine points, but correlating specs with the physical driver is very important. The specs at best are a very indirect pointer to the complexity of the real-world system. The real values of specs is telling us something about how the driver is made - this, by the way, is why I like to see drivers tested on large flat baffles, so I can see the driver, not the baffle artifacts. The "differences" between the 15MB700 and 15NMB420 could be nothing more than different-sized baffles - indeed, baffle diffraction alone could account for the ripples I complained about in the Supravox CSD measurement.
Lynn Olson said:
The Supravox just has much less internal loss (because of its low mass relative to the B&C driver) and as a result has less control about 20 db down. Note that this is likely a driver that would respond well to Bud's Enabl pattern, especially considering that the surround likely provides little in the way of edge damping.
However the "wavy'ness" for the first 5 db "down" at higher freq.s is likely a sign of cone surface loading.
I'd personally opt for the 165 GMF - that lower mass is quite noticeable when it comes to perception of detail.
Still though, the Audax PR17M0 if used properly provides a lot more value than most other offerings, and is still fairly low in mass. But frankly I like your idea of the highest eff. 18 Sound 6.5 and 10 inch driver pairing - THAT should be extremely detailed and worth the price premium.
ScottG said:
Yes, I agree that the lower-mass driver, all other things being equal, being the best choice for resolution - with the caution that lower-mass drivers may also break up at a lower SPL level than a heavier, stiffer cone. However, with the efficiency of prosound being what it is, the onset of cone breakup is probably at an extremely high SPL - we're not talking the ~95 DB SPL of a direct-radiator Lowther, AER, or Feastrex here.
The combo of the 18Sound 6ND410 (EBP = 444) and 10NDA610 (EBP = 371), with 10NDA610 starting an early rolloff around 1k Hz and the 6ND410 rolling off later at 3 kHz, and sharing a common electrical highpass around 300~400 Hz, should assure very high SPL's at low distortion. The low-Q (0.5 or less) 2nd-order electrical highpass would match the baffle-peak frequency, forming an acoustic 12~18 dB/octave highpass that would very gradually steepen to 30 dB/octave below 45 Hz (1/2 the Fs of the open-baffle 10NDA610).
The complement to the acoustic highpass around 300 Hz would be the electrical lowpass of the bass module, which I plan to be a quartet of 15" drivers with a slight 15~20 degree crease across the front panel, giving a slight "V" to the plan view (as seen from the top). The JBL Everest system, with its side-by-side 15" drivers in a slight "V" profile, had the clever idea of separate low-pass filters for the Inside and Outside woofers, which gave better lateral dispersion than a common lowpass filter for both drivers. This could apply to the bass module as well, with separate lowpass filters for the Inside and Outside pairs of drivers.
Suggestions for medium-to-high Qts, good-sounding 15" bass drivers (neodymium magnets preferred) are welcome. There's always Eminence, I guess.
The only one I've seen is what John is building (custom)
over at AE speakers:
AE speakers forum
And an old page that describes it:
Dipole 15
But it's not priced like the eminence!
herm
Edit: I've never heard it - so no opinion on how it sounds...
There isn't much.
I would hesitate to say that the AE speakers are current production. I'm not trying to cause trouble, I just don't know about long term availability.
http://www.hometheatershack.com/forums/fi-audio/7153-15-ib-sub.html
is another option. The Re is really low and if you were thinking about paralleling a bunch to raise the sensitivity I don't think it is a go.
Dayton makes an IB woofer.
Kevin Haskins at DIYCable has an OB 10" cooking, and a few larger woofers with medium Qt.
GR Research/Rhythmik has 12" OB woofers (servo/active) coming in early 08.
The last thing I can think of is contracting with Skaaning/Audio Technology to make exactly what you want.....$$$
C
I would hesitate to say that the AE speakers are current production. I'm not trying to cause trouble, I just don't know about long term availability.
http://www.hometheatershack.com/forums/fi-audio/7153-15-ib-sub.html
is another option. The Re is really low and if you were thinking about paralleling a bunch to raise the sensitivity I don't think it is a go.
Dayton makes an IB woofer.
Kevin Haskins at DIYCable has an OB 10" cooking, and a few larger woofers with medium Qt.
GR Research/Rhythmik has 12" OB woofers (servo/active) coming in early 08.
The last thing I can think of is contracting with Skaaning/Audio Technology to make exactly what you want.....$$$
C
I mustn't be as silly as I thought, because I was contemplating the 6ND410 + 10NDA610 combo myself last night.Lynn Olson said:
For the 15's, the only high Qts pro drivers I can think of that fits your spec is the 18Sound 15W500, and the Eminence Delta 15LF and Deltalite 2515. The last is neo, the others ceramic.
Edit: as you're going to use SS for the LF is it possible to use a variable output impedance system like
http://sound.westhost.com/project56.htm
or one of Nelson's designs provided the power is adequate?
I chose the Delta 15LF for the DarkStar, and in their double Ripol alignment they make for a very powerful low end. I can offer a first hand recommendation. US Speaker was very good at getting 20 of them to me. I personally wouldnt look any further if I were contemplating 15's in this way.
Lynn Olson said:
This is one time where I would definitely start looking hard at a driver based on your own specifications.
Eminence is the natural suggestion, but..
Ciare in general has not only the ability to create a better driver, but also requires only 20 drivers per order.
I have no idea what either 18 Sound or BMS Pro requires (or even if they would produce a specialty or "OEM" driver), of course they would be worth checking on - particularly 18 Sound if the 6.5 and 10 inch driver combo works out.
There is also Lyeco in Taiwan.
IMO it really needs 2 very different drivers:
1: A low mass VERY high eff. midbass (about 104 db) driver with low distortion and an fs near 60 Hz and a qts right near .7 - all with a linear higher freq. response and decent excursion (say +/- 4mm). To the best of my knowledge no one produces such a driver in that the driver's vas would be ludicrous. The Jensen Neo 150 would be a poor but almost tenable substitute given its raised response near 140 Hz.
2: The other *3* drivers should be lower freq. extended drivers and should cross around that 70 Hz region. They should have a summed impedance (when connected in parallel) that gives you the substantial spl's you will need. Qts (around 1.4) and excursion should be high (11-14 mms). Fs should be low - about 25 Hz to stay out of the range of the majority of fundamentals. I can't think of any production driver that meets these requirements (and of course vas of such a driver would be beyond ludicrous).
Barring the above..
a 2 driver "Everest" solution then:
inside driver is an alpha 15, outside driver is a Jensen Neo 150.
BTW, one of my favorite posted threads is this trip to Japan which happens to be primarily for the Everest debut (day 2 has info on the Everest):
http://audioheritage.org/vbulletin/showthread.php?t=12157
ScottG said:
Well, I've been pondering the various ways a bass module would work. After hearing the OB systems at the RMAF, the "audiophile" flavors of OB are laughably short of excursion and dynamics. I know very well the sound of overstressed drivers and I'm amazed these big-name guys are getting away with it - audiences are startled by the "no-box" sound and don't notice how stressed the drivers are. Well, I guess the audiophile OB's are less stressed than an electrostat or magnetic-planar, but that doesn't say much, does it?
The pair of 15's I heard in the Emerald Physics room struck me as the minimum for reasonable dynamics in the bass/midbass region - and that was in a very small hotel room. It was nowhere close to what a well-engineered bass horn can do, but did have better dynamics than a generic audiophile closed-box speaker. The hand of heavy digital EQ was pretty obvious, though - it only reinforced my feeling that increased driver area is a far better solution than EQ and lots of transistor watts heating up the voice-coils.
Thus, my determination to use 4X 12", 2X 12" + 2X 15", or 4X 15". The quartet of drivers would be connected in series-parallel where one arm is fairly wideband (60 ~ 300 Hz passband, probably with low-to-medium Qts and using advanced pole-piece construction to lower IM distortion) and the other arm is a brute-force large-area bass driver (60 ~ 120 Hz passband and with higher Qts). The wider-range drivers can either be on the top row or on the "inner" side, a la JBL Everest. You can make a good case for either method, particularly if the wider-range drivers are 12" instead of 15".
As for bi-amping vs a single amp, it would be nice if the bass module had enough efficiency and frequency-shaping capability that single-amping remains an option. Bi-amping does offer the increased flexibility of a Z-matic style adjustable damping factor (easy to do with any feedback amplifier) as well as precise Left and Right channel room compensation.
I am still amazed that serious loudspeaker designers have been able to fool themselves that bass distortion doesn't matter that much, that small-signal EQ is some kind of solution to the dipole rolloff problem. This is the same kind of thinking behind the B*** 901 - c'mon guys, we all know the limits of EQ by now, increasing amp power to overcome loss of efficiency is the shortest path to restricted headroom and increased IM distortion.
P.S. Thanks for all the info and suggestions, much appreciated!!! The links brought back memories of Japan (1956-61) - sure wish we had that kind of slick, fast, and wonderful all-electric train system here in the USA! Instead, all we get are airlines with creaky 30-year-old airplanes and intrusive "homeland security" checks. I'm amazed anyone flies at all anymore, it's so degrading.