No reason it shouldn't work. Might even be a bit better, assuming the right midbass driver. The Beyond the Ariel project was always centered around a large-format 2-way ... 15" Altec-style driver, large-format compression driver (3" diaphragm), a low-diffraction horn with a mouth area comparable to the 15" driver, and a crossover in the 600 to 800 Hz region. That is the upper size limit for a 2-way, just as the lower size limit for a 2-way (that isn't a desktop speaker) is a 6.5" midbass and a 1" dome tweeter.
My biggest gripe about open baffles is the lack of power-handling at low frequencies, combined with the need for below-baffle-peak equalization, problems a well-designed closed-box with a decent woofer doesn't have. You've solved that problem by intelligently combining the two, with identical drivers and a passive series crossover at an appropriate frequency. (One variation might be a 2nd-order series crossover, but that introduces a 180-degree allpass phase rotation at a frequency where it is probably audible. So probably not.)
A hybrid with a closed-box and OB on top could have very good impulse response, without box-induced hangover and resonance in the critical 200 to 800 Hz range. No issues with pipe modes in the cylindrical vent, or the subtler annoyance of whether to place the vent on the front, back, or sides of the enclosure.
I hate to suggest this ... but that won't stop me. You could do a Mark II of the Bitches Brew using this configuration ... 15" closed-box bass, 15" open baffle midbass (same driver), and a large format horn. Try the AH425 Azurahorn, which has a 16" diameter. Might be interesting.
My biggest gripe about open baffles is the lack of power-handling at low frequencies, combined with the need for below-baffle-peak equalization, problems a well-designed closed-box with a decent woofer doesn't have. You've solved that problem by intelligently combining the two, with identical drivers and a passive series crossover at an appropriate frequency. (One variation might be a 2nd-order series crossover, but that introduces a 180-degree allpass phase rotation at a frequency where it is probably audible. So probably not.)
A hybrid with a closed-box and OB on top could have very good impulse response, without box-induced hangover and resonance in the critical 200 to 800 Hz range. No issues with pipe modes in the cylindrical vent, or the subtler annoyance of whether to place the vent on the front, back, or sides of the enclosure.
I hate to suggest this ... but that won't stop me. You could do a Mark II of the Bitches Brew using this configuration ... 15" closed-box bass, 15" open baffle midbass (same driver), and a large format horn. Try the AH425 Azurahorn, which has a 16" diameter. Might be interesting.
Last edited:
Crossing monopole and dipole woofers (and getting cardioid response around transition) is nothing new. To my knowledge Jorma Salmi of Gradient introduced it in his Gradient 1.2 model and successors. I copied this in my AINOgradients.
John Kreskowsky tells more about this here http://musicanddesign.speakerdesign.net/craw_cross.html
John Kreskowsky tells more about this here http://musicanddesign.speakerdesign.net/craw_cross.html
Lynn,
That is a GREAT idea. Somehow I had missed the Azurahorn people. What a... well, quaint, but also fantastic website. Stunning measurements from those drivers. Bitches Brew Mark II: Bottom driver, 15" sealed, middle section Azurahorn, top section 15" dipole woofer. Lots to explore there.
Not sure if I have room for a BB2 setup, which is also similar to what I've published in the Beyond the Ariel thread elsewhere on this forum. The speaker I'm planning to build with my neighbor, Thom Mackris of Galibier Designs, is a 3.9 cubic foot/110 liter low-diffraction (4" radius rounds on left and right edges) closed box with a Great Plains Audio 416 Alnico magnet 15" driver, and a Athos Audio Yuichi A290 wood horn with an adapter for a 1.4" exit compression driver, like the 18Sound ND3SN with a 3" titanium-nitride diaphragm.
Since the Yuichi A290 has a near-flat response, not quite as ruler-flat as the AH425 Azurahorn, but still quite good, crossover should be simple with just a bit of shelf equalization in the HF.
Seeing the article you published in this thread made me realize what you've come up could be generalized to any 2-way, but is particularly apt for large-format 2-ways, like mine, or any other large-format monitor speaker. At the expense of some additional height and depth (keeping the same 110 liter volume for the closed box), your concept would work just fine for the Beyond the Ariel, taking it to the next level beyond a large-format 2-way.
The new open-baffle midsection would retain the 4" radius quarter-rounds, and might have little winglets for structural and visual reasons, but the LF/MF crossover would still fall in the 600 to 800 Hz region, and overall system design would be quite similar. I could save a bit of expense by reserving the Alnico woofer for the upper section, and select a much cheaper ceramic magnet 416 for the closed box (the two versions are otherwise the same).
Here's a drawing of what I have so far, and how an OB midsection might be easily added to it, between the bass bin and the wood horn. From preliminary calculations on a bass bin with less height but more depth (but the same 26" width), the overall system would gain another 8" of height, to accommodate the stacked 16" frames of the GPA/Altec 416 drivers.
Since the Yuichi A290 has a near-flat response, not quite as ruler-flat as the AH425 Azurahorn, but still quite good, crossover should be simple with just a bit of shelf equalization in the HF.
Seeing the article you published in this thread made me realize what you've come up could be generalized to any 2-way, but is particularly apt for large-format 2-ways, like mine, or any other large-format monitor speaker. At the expense of some additional height and depth (keeping the same 110 liter volume for the closed box), your concept would work just fine for the Beyond the Ariel, taking it to the next level beyond a large-format 2-way.
The new open-baffle midsection would retain the 4" radius quarter-rounds, and might have little winglets for structural and visual reasons, but the LF/MF crossover would still fall in the 600 to 800 Hz region, and overall system design would be quite similar. I could save a bit of expense by reserving the Alnico woofer for the upper section, and select a much cheaper ceramic magnet 416 for the closed box (the two versions are otherwise the same).
Here's a drawing of what I have so far, and how an OB midsection might be easily added to it, between the bass bin and the wood horn. From preliminary calculations on a bass bin with less height but more depth (but the same 26" width), the overall system would gain another 8" of height, to accommodate the stacked 16" frames of the GPA/Altec 416 drivers.
Attachments
Last edited:
Overall sounds like an excellent strategy. I get the basic idea of the 416’s, but I’m not intimately familiar with the history. It doesn’t seem like a driver designed for acoustic suspension and I would be inclined to think that for a subwoofer, a lot of other drivers would be better. What do the Great Plains drivers do that other units don’t?
I used to think all drivers sounded the same in the piston-band (flat-response region), but I was wrong. In particular, I compared a JBL 2235, which is a modern, well-regarded driver, against an Alnico-magnet 416 (current-production GPA). To my great surprise, they sounded very different, and nothing alike. And this wasn't at head-banging levels, or trying to get thumping sub-bass out of them. And the difference wasn't so much in the deep bass, but in the lower midrange and upper bass.
Eventually I determined it must have had something to do with the magnetic construction of the two drivers. One was overhung in the modern style, and had a shaded-pole construction to control the fringing flux at the edge of the gap. The other was underhung, so no VC was ever out of the gap, and the field intensity in the gap was high enough to partially saturate the pole pieces.
In any event, the Altec/GPA has a distinct sound, but it's not distortion, and it's not peaks, either. I measured it and it is very smooth, more so than modern prosound drivers, especially at the top of the passband. So it went to the top of the list, along with the fabled TAD Alnico-magnet drivers. But they are very very expensive.
Altec used the 803/416 in large vented boxes, ranging in size from the 12.5 cubic feet of the 9845 monitor, to about 10 cubic feet in the Valencia and Model 19. The Model 19 was the only post-Theile/Small system; the other two precede T/S, coming out in the late Sixties. At any rate, the current version is suitable for closed boxes of 3.5 to 4 cubic feet, and vented boxes of 8 to 10 cubic feet ... however, parameters (cone mass and BL product) of the 416 varied quite a lot over the decades since the early Fifties, so there is no "one" 416, so no surprise the Altec cabinets were all over the place, and by modern standards, very crude.
This is particularly noticeable when the Altec/GPA drivers are taken out of the big flappy unreinforced plywood cabinets that Altec used back in the day. Those cabinets are massively colored, but so were other contemporary cabinets from JBL, Klipsch, Jensen, et al. They were no better, just big plywood boxes. It was the Brits who were the pioneers in controlling cabinet coloration, not the Americans. It takes a modern cabinet to hear what the Altec drivers really sound like, which is very fast and crisp, and bright, vivid tone colors ... the famous West Coast Sound, but without the unpleasant upper-bass and lower-midrange colorations (thank the cheap cabinets for those).
Of course, there are lots and lots of modern 15" prosound drivers. Most are optimized for very heavy-duty sound reinforcement use, others for movie theaters, and only a few are optimized for wideband midbass use in large-format studio monitors. They tend to have lighter cones (less than 80 grams), smaller voice coils (3 inches instead of 4), and typically have smoother responses above 1.5 kHz. The big boys designed for subwoofer use have insanely high power ratings, very large voice coils, very heavy cones, and response that is garbage above 500 Hz. Horses for courses, as the Brits say.
The selection of the GPA 416 is a subjective choice, based on vivid tone colors, the hallmark of this particular driver. I'm sure there are other choices, but I'm not going to plow through scores of them on a subjective basis. They do sound different, against all expectation, even though they are used in the flat-passband region. I'm sure the listeners of this forum have their favorites.
In a modern context, the 416 is a midbass driver, with very smooth response, and with only moderate power-handling, 60 watts or less, because of the small and light voice coil. In an age of 200 to 500 watt Class D amps, it's not a suitable choice for many. But for folks with 8 to 60-watt tube amps, it's at the top of the list, along with TAD and a few others, including exotica like the 15" field-coil drivers from Kanzen and Wolf von Langa in Germany.
So there are many choices, from entry-level Eminence to Wolf von Langa's really over-the-top drivers, all with 15" frames. That's one of the charms of a large-format 2-way ... stupendous dynamic range, no longer limited by the typical audiophile setup with multiple 8 to 10-inch woofers and an MTM around an exotic dome tweeter or ribbon. That's a legitimate choice, but the sound is completely different than a large-format 2-way system. As the V8 fans say, ain't no replacement for displacement. And I agree.
I'm not sure what to call your new design ... it's very close to a classical 2-way, and not like the bandsplitting of a 3-way at all. Nor is it like a 2.5 way, with bass augmentation for a midbass driver that's come up a bit short in the depth and power department.
It changes the basic character of the midbass driver, freeing it from box coloration, while retaining its power where it matters, below 100 Hz. That's a big deal, a really big deal. And the choice of identical drivers for the entire bass range gets away from the challenging task of tone-matching dissimilar drivers, which is the downfall of so many 3-way systems. Many 3-ways suffer from disjointed sound as playback level is changed. They sound great from 70 to 85 dB average SPL, but come apart above and below that. A good 2-way can get around that, if dynamic range is sufficient.
Eventually I determined it must have had something to do with the magnetic construction of the two drivers. One was overhung in the modern style, and had a shaded-pole construction to control the fringing flux at the edge of the gap. The other was underhung, so no VC was ever out of the gap, and the field intensity in the gap was high enough to partially saturate the pole pieces.
In any event, the Altec/GPA has a distinct sound, but it's not distortion, and it's not peaks, either. I measured it and it is very smooth, more so than modern prosound drivers, especially at the top of the passband. So it went to the top of the list, along with the fabled TAD Alnico-magnet drivers. But they are very very expensive.
Altec used the 803/416 in large vented boxes, ranging in size from the 12.5 cubic feet of the 9845 monitor, to about 10 cubic feet in the Valencia and Model 19. The Model 19 was the only post-Theile/Small system; the other two precede T/S, coming out in the late Sixties. At any rate, the current version is suitable for closed boxes of 3.5 to 4 cubic feet, and vented boxes of 8 to 10 cubic feet ... however, parameters (cone mass and BL product) of the 416 varied quite a lot over the decades since the early Fifties, so there is no "one" 416, so no surprise the Altec cabinets were all over the place, and by modern standards, very crude.
This is particularly noticeable when the Altec/GPA drivers are taken out of the big flappy unreinforced plywood cabinets that Altec used back in the day. Those cabinets are massively colored, but so were other contemporary cabinets from JBL, Klipsch, Jensen, et al. They were no better, just big plywood boxes. It was the Brits who were the pioneers in controlling cabinet coloration, not the Americans. It takes a modern cabinet to hear what the Altec drivers really sound like, which is very fast and crisp, and bright, vivid tone colors ... the famous West Coast Sound, but without the unpleasant upper-bass and lower-midrange colorations (thank the cheap cabinets for those).
Of course, there are lots and lots of modern 15" prosound drivers. Most are optimized for very heavy-duty sound reinforcement use, others for movie theaters, and only a few are optimized for wideband midbass use in large-format studio monitors. They tend to have lighter cones (less than 80 grams), smaller voice coils (3 inches instead of 4), and typically have smoother responses above 1.5 kHz. The big boys designed for subwoofer use have insanely high power ratings, very large voice coils, very heavy cones, and response that is garbage above 500 Hz. Horses for courses, as the Brits say.
The selection of the GPA 416 is a subjective choice, based on vivid tone colors, the hallmark of this particular driver. I'm sure there are other choices, but I'm not going to plow through scores of them on a subjective basis. They do sound different, against all expectation, even though they are used in the flat-passband region. I'm sure the listeners of this forum have their favorites.
In a modern context, the 416 is a midbass driver, with very smooth response, and with only moderate power-handling, 60 watts or less, because of the small and light voice coil. In an age of 200 to 500 watt Class D amps, it's not a suitable choice for many. But for folks with 8 to 60-watt tube amps, it's at the top of the list, along with TAD and a few others, including exotica like the 15" field-coil drivers from Kanzen and Wolf von Langa in Germany.
So there are many choices, from entry-level Eminence to Wolf von Langa's really over-the-top drivers, all with 15" frames. That's one of the charms of a large-format 2-way ... stupendous dynamic range, no longer limited by the typical audiophile setup with multiple 8 to 10-inch woofers and an MTM around an exotic dome tweeter or ribbon. That's a legitimate choice, but the sound is completely different than a large-format 2-way system. As the V8 fans say, ain't no replacement for displacement. And I agree.
I'm not sure what to call your new design ... it's very close to a classical 2-way, and not like the bandsplitting of a 3-way at all. Nor is it like a 2.5 way, with bass augmentation for a midbass driver that's come up a bit short in the depth and power department.
It changes the basic character of the midbass driver, freeing it from box coloration, while retaining its power where it matters, below 100 Hz. That's a big deal, a really big deal. And the choice of identical drivers for the entire bass range gets away from the challenging task of tone-matching dissimilar drivers, which is the downfall of so many 3-way systems. Many 3-ways suffer from disjointed sound as playback level is changed. They sound great from 70 to 85 dB average SPL, but come apart above and below that. A good 2-way can get around that, if dynamic range is sufficient.
Last edited:
Let's waste a few more electrons talking about air loads. As anyone who's T/S measured a driver knows, the air load of the driver in the test box, and hanging in free air, are not the same. It's a correction factor you have to crank into the equations ... typically not much, a few grams of equivalent cone mass, somewhere around 10% or so.
This air load is actually the coupling to air, by which the driver radiates sound into the room. And it changes, depending in proximity to other objects. In free air, hanging from a string, it is at a minimum. Put the driver in a well-sealed box, it goes up (by a few grams). Put the box next to another box with an active driver, and it goes up some more. Put the box on a hard surface, and the floor image does the same thing. Put the box at the small end of a bass horn, and it goes up even more.
Notice the pattern? As the air load increases, adding a few grams each time, the efficiency goes up. By several times by the time we get to a bass horn.
What does this arcane information have to do with speaker design? Not much, most of the time. But ... there's always that pesky floor reflection, which behaves just like a mirror-image speaker just below the floor, with its own amplifier. At long wavelengths, this floor image is in-phase with the real, physical speaker, and acts just like we've stacked two speakers on top of each other.
I think you can see the problem approaching. The two woofers, real and virtual, are close to each other, and the air load goes up, doubling their efficiency (and doubling the power-handling, too). What's the problem? Unfortunately, the two tweeters are many wavelengths apart, and phase addition is random. At some frequencies, they are out of phase, at others, ninety degrees apart, and others, in-phase. This is annoying when trying to measure speakers, and the floor reflection is typically gated out.
I think you can now see what's happening with those huge WMTMW systems you see and hear at shows. Ever wonder why they sound so disjointed and incoherent? The air load for the upper woofer is pretty much in free air, far way from the floor, while the lower woofer is closely coupled with the floor image. They are operating at different efficiencies, and not surprisingly, sound different. You'd expect with different effective cone masses they'd sound different, and they do. That's what you're hearing. You're not imagining it, what you're hearing is real.
What about the hybrid speaker? True, the same effect is happening here, but two things are making a big difference: the upper and lower cones are closer together, and more critically, They Are The Same. This is the secret. Dissimilar air loads are bad enough, but dissimilar cones, with their own distinct resonances and nonlinearities, are a disaster, when added to different air loads. Think of those four cones in the WMTMW stack; in reality, all four are different from each other!
So there is no wonderful efficiency gain with the hybrid, unlike simply connecting the two bass drivers in parallel. But that's a good thing, not bad. No comb-filtering from four effective drivers, when counting the floor image. No weird effects from dissimilar cones with different air loads, unlike a massive WMTMW. Just two nearly identical cones with only slightly different air loads (closed box vs open baffle).
This air load is actually the coupling to air, by which the driver radiates sound into the room. And it changes, depending in proximity to other objects. In free air, hanging from a string, it is at a minimum. Put the driver in a well-sealed box, it goes up (by a few grams). Put the box next to another box with an active driver, and it goes up some more. Put the box on a hard surface, and the floor image does the same thing. Put the box at the small end of a bass horn, and it goes up even more.
Notice the pattern? As the air load increases, adding a few grams each time, the efficiency goes up. By several times by the time we get to a bass horn.
What does this arcane information have to do with speaker design? Not much, most of the time. But ... there's always that pesky floor reflection, which behaves just like a mirror-image speaker just below the floor, with its own amplifier. At long wavelengths, this floor image is in-phase with the real, physical speaker, and acts just like we've stacked two speakers on top of each other.
I think you can see the problem approaching. The two woofers, real and virtual, are close to each other, and the air load goes up, doubling their efficiency (and doubling the power-handling, too). What's the problem? Unfortunately, the two tweeters are many wavelengths apart, and phase addition is random. At some frequencies, they are out of phase, at others, ninety degrees apart, and others, in-phase. This is annoying when trying to measure speakers, and the floor reflection is typically gated out.
I think you can now see what's happening with those huge WMTMW systems you see and hear at shows. Ever wonder why they sound so disjointed and incoherent? The air load for the upper woofer is pretty much in free air, far way from the floor, while the lower woofer is closely coupled with the floor image. They are operating at different efficiencies, and not surprisingly, sound different. You'd expect with different effective cone masses they'd sound different, and they do. That's what you're hearing. You're not imagining it, what you're hearing is real.
What about the hybrid speaker? True, the same effect is happening here, but two things are making a big difference: the upper and lower cones are closer together, and more critically, They Are The Same. This is the secret. Dissimilar air loads are bad enough, but dissimilar cones, with their own distinct resonances and nonlinearities, are a disaster, when added to different air loads. Think of those four cones in the WMTMW stack; in reality, all four are different from each other!
So there is no wonderful efficiency gain with the hybrid, unlike simply connecting the two bass drivers in parallel. But that's a good thing, not bad. No comb-filtering from four effective drivers, when counting the floor image. No weird effects from dissimilar cones with different air loads, unlike a massive WMTMW. Just two nearly identical cones with only slightly different air loads (closed box vs open baffle).
Last edited:
@Lynn Olson Fascinating discussion from you about the Altecs. So permit me me say back to you in my own language, what I believe you said to me:
Because of a whole catalog of 100 year old baked-in assumptions about speaker design, including things like "Amps are always voltage sources" (not necessarily) and "crossovers are passive" (not necessarily) and "signal from the amp is not EQ'd before getting to the speaker" and "response of a speaker is mass-controlled above resonance" and "the ultimate goal of every speaker is flat frequency response" and "woofers HAVE to have a box" we box ourselves in to all kinds of assumptions that CAN be discarded.
If DSP EQ is available before the signal gets to the amp, at least half of those assumptions can be tossed out the window. In the old paradigm you always trade SPL against low frequency cutoff. With DSP you don't have to do that.
I think high-xmax low efficiency woofers with long heavy voice coils and heavy cones measure great but still sound like lifeless un-musical thumps.
I theorize that efficiency is a big part of that. There is an essential virtue of efficiency in that the source is being connected to the air in the room more effectively; and when that number is high you retain some organic details of the signal that low efficiency systems lose.
So if you want a sealed box subwoofer that's flat down to 25Hz, you can buy a 15" sub with 86dB SPL and 14mm Xmax. It will cost you $200. You put it in a sealed and get an F3 of 25Hz. Part of how you got that low F3 is you used a smallish magnet to raise the Q.
But there are other ways to slay that beast.
You could also buy a 15" sub with >/= 96dB SPL and >/= 14mm Xmax. If you search loudspeakerdatabase.com you'll find a very few "needle in haystack" drivers that do that:
http://www.loudspeakerdatabase.com/search/14.0_size_in_16.0/95.9_spl_112.0/13.9_xmax_42.5
And they all cost $500 to $1000.
Now what you have is a high SPL low mass driver with low Q. And to most speaker builders who live in the baked-in assumptions of the 1950s and Acoustic Research, it's a lousy choice because it appears to need a sealed box of 30 liters and an F3 of 70Hz.
But that's not a problem, because you're no longer playing a game of "specsmanship" as though the 70Hz figure really matters. You can still put it in a 50 or 100 liter box, and the bigger box WILL buy you more efficiency below resonance, even if it doesn't push the resonant frequency down all that much. And you can add EQ to make up the difference. So your speaker is still 83dB at 25Hz (just like the Dayton subwoofer) but it's more efficient at all points above 25Hz - it's 96dB at 70Hz and up. It has lower mass, couples to the air better, and blends better with midrange drivers.
The problem with all those low SPL monster subs with 1000 watt amps is they never blend well with any satellites. They're impressive but they still sound like thumps.
[The thing I like about the 15OB350 from SB Acoustics is, they've managed to make a high efficiency driver that also has a higher Q because they designed a lightweight diaphragm to have a large Vas which is how you squeeze the most juice out of an Open Baffle. I suspect this driver will also do well in a large sealed box.]
My efficiency theory is why I chose the 96dB Dayton PA310 for my Flanagangsters. It's got 5mm xmax which for a box tuned to 42 Hz is plenty but it's high SPL and it's a really nice midrange. I am prioritizing high frequency bandwidth over low frequency heft - and I'm using the DSP to protect it from signals below 35Hz so excursion is minimized. Thus I get the "aliveness" of a high efficiency design and still get the low frequency output that I want.
What I think I hear you saying is that the Altec underhung woofers are high efficiency; and since you're not a high SPL "flap my pantlegs with bass" guy you're willing to trade away some Xmax to get some midband aliveness and color and high end extension.
I'm wondering if the drivers in that LoudSpeakerDatabase search link above will give you the same color and responsiveness that your ears tell you the low SPL drivers are missing.
~
Now there is also an issue of the Altecs having Alnico magnets. Alnico was starting to disappear just when I started building speakers in the 80s so I was never part of that scene. What I pick up is that a lot of people just like the sound of Alnico. Sort of like some people like the sound of tubes or MOSFETS or whatever.
So I have no personal experience with this but I do have a theory.
When I worked at Jensen I took a trip to Fujicone, a cone manufacturer owned by Jensen. They made the cones for the Bose 901s and for Jensen who made 70,000 speakers a day. Visiting cone factories is like a strange voyage into an anachronistic witches brew cauldron of slurries and chemical paste and steam going everywhere... you almost feel like you're walking into the 1400s. You half expect a figure draped in black to come out of the steam with a crystal ball.
I talked to cone manufacturers all the time. You would call them up and say "Hey George I ran the curves on these speakers and I need more high end / I need a thicker cone / I want to try a felted cone instead of a pressed cone" or whatever.
They would cook something up for you and send you a few dozen sample cones, then rinse repeat until you got what you wanted.
I was puzzled as to how these guys had any objective sense of what they were doing. Was it all just witchcraft? Or did they have measurement equipment?
The chief engineer was this guy named George Pope. I asked him, "George do you have some speaker-driving fixture around here, that you stick all these cones on so you can run your own tests?"
"Nope. Here, let me show you how I evaluate a speaker cone." He walks us into the bathroom.
He says "This bathroom is super reverberant, it echoes for half a second when you snap your fingers. I just walk in here with a speaker cone and I flick the edge of the cone. A hard pressed paper cone makes a sharp crisp cracking noise. A soft felted cone makes a dull cardboardy sound. A poly cone makes a plastic-y sound. A metal cone makes a metallic sound. I don't measure anything. I just snap the edge of the cone in the men's room and listen to the echo. That tells me exactly what the cone is going to sound like."
Dang was that ever an epiphany for me. Because George flicking the cone is an impulse; and what you hear with your ears is an FFT.
That cone material has an intrinsic signature sound that is independent of the shape or size of the cone or whatever else it's attached to. It has its own signature sound.
From that point forward it was obvious to me that pressed cones have a certain sound, felted cones have a certain sound, dense vs. not dense, plastic vs paper vs metal. Same with tweeter domes. Same with compression driver diaphragms.
Same with box materials. Particleboard does not sound like plywood which does not sound like injection molded plastic which does not sound like marble which does not sound like machined aluminum.
Today in the age of DSP when you can EQ anything "ruler flat" with the touch of a button, you still cannot make a paper cone tweeter sound like a soft dome tweeter, which does not sound like titanium which does not sound like aluminum which does not sound like beryllium.
I'm pretty sure you can measure some of those differences which you so easily hear, but they do not easily or readily show up in the measurements. They may not show up at all in the frequency response curves. I suspect there are characteristics which you would be hard pressed to define purely in terms of measurements.
So I can't help but wonder if Alnico likewise has its own signature sound. How can it not? It has very peculiar hysteresis / saturation curves etc.
Ditto ferrite.
Ditto neodymium.
Materials have their own signature sound, no matter how they are used. Right now we're talking about magnetics.
How about BJTs vs FETs vs JFETS vs class A vs AB vs various class D implementations? Or triodes or pentodes?
Just like EQ can't make a metal dome sound like a soft dome, maybe 40 dB of negative feedback can't make a BJT sound like a MOSFET either. You can get close but there is always a "grain" in the sound that the trained ear can detect.
The "objective measurement people" are always skeptical about this, but the subjective people consistently maintain their preferences over long periods of time.
Maybe Alnico has some secret sauce that ferrite doesn't have
Because of a whole catalog of 100 year old baked-in assumptions about speaker design, including things like "Amps are always voltage sources" (not necessarily) and "crossovers are passive" (not necessarily) and "signal from the amp is not EQ'd before getting to the speaker" and "response of a speaker is mass-controlled above resonance" and "the ultimate goal of every speaker is flat frequency response" and "woofers HAVE to have a box" we box ourselves in to all kinds of assumptions that CAN be discarded.
If DSP EQ is available before the signal gets to the amp, at least half of those assumptions can be tossed out the window. In the old paradigm you always trade SPL against low frequency cutoff. With DSP you don't have to do that.
I think high-xmax low efficiency woofers with long heavy voice coils and heavy cones measure great but still sound like lifeless un-musical thumps.
I theorize that efficiency is a big part of that. There is an essential virtue of efficiency in that the source is being connected to the air in the room more effectively; and when that number is high you retain some organic details of the signal that low efficiency systems lose.
So if you want a sealed box subwoofer that's flat down to 25Hz, you can buy a 15" sub with 86dB SPL and 14mm Xmax. It will cost you $200. You put it in a sealed and get an F3 of 25Hz. Part of how you got that low F3 is you used a smallish magnet to raise the Q.
But there are other ways to slay that beast.
You could also buy a 15" sub with >/= 96dB SPL and >/= 14mm Xmax. If you search loudspeakerdatabase.com you'll find a very few "needle in haystack" drivers that do that:
http://www.loudspeakerdatabase.com/search/14.0_size_in_16.0/95.9_spl_112.0/13.9_xmax_42.5
And they all cost $500 to $1000.
Now what you have is a high SPL low mass driver with low Q. And to most speaker builders who live in the baked-in assumptions of the 1950s and Acoustic Research, it's a lousy choice because it appears to need a sealed box of 30 liters and an F3 of 70Hz.
But that's not a problem, because you're no longer playing a game of "specsmanship" as though the 70Hz figure really matters. You can still put it in a 50 or 100 liter box, and the bigger box WILL buy you more efficiency below resonance, even if it doesn't push the resonant frequency down all that much. And you can add EQ to make up the difference. So your speaker is still 83dB at 25Hz (just like the Dayton subwoofer) but it's more efficient at all points above 25Hz - it's 96dB at 70Hz and up. It has lower mass, couples to the air better, and blends better with midrange drivers.
The problem with all those low SPL monster subs with 1000 watt amps is they never blend well with any satellites. They're impressive but they still sound like thumps.
[The thing I like about the 15OB350 from SB Acoustics is, they've managed to make a high efficiency driver that also has a higher Q because they designed a lightweight diaphragm to have a large Vas which is how you squeeze the most juice out of an Open Baffle. I suspect this driver will also do well in a large sealed box.]
My efficiency theory is why I chose the 96dB Dayton PA310 for my Flanagangsters. It's got 5mm xmax which for a box tuned to 42 Hz is plenty but it's high SPL and it's a really nice midrange. I am prioritizing high frequency bandwidth over low frequency heft - and I'm using the DSP to protect it from signals below 35Hz so excursion is minimized. Thus I get the "aliveness" of a high efficiency design and still get the low frequency output that I want.
What I think I hear you saying is that the Altec underhung woofers are high efficiency; and since you're not a high SPL "flap my pantlegs with bass" guy you're willing to trade away some Xmax to get some midband aliveness and color and high end extension.
I'm wondering if the drivers in that LoudSpeakerDatabase search link above will give you the same color and responsiveness that your ears tell you the low SPL drivers are missing.
~
Now there is also an issue of the Altecs having Alnico magnets. Alnico was starting to disappear just when I started building speakers in the 80s so I was never part of that scene. What I pick up is that a lot of people just like the sound of Alnico. Sort of like some people like the sound of tubes or MOSFETS or whatever.
So I have no personal experience with this but I do have a theory.
When I worked at Jensen I took a trip to Fujicone, a cone manufacturer owned by Jensen. They made the cones for the Bose 901s and for Jensen who made 70,000 speakers a day. Visiting cone factories is like a strange voyage into an anachronistic witches brew cauldron of slurries and chemical paste and steam going everywhere... you almost feel like you're walking into the 1400s. You half expect a figure draped in black to come out of the steam with a crystal ball.
I talked to cone manufacturers all the time. You would call them up and say "Hey George I ran the curves on these speakers and I need more high end / I need a thicker cone / I want to try a felted cone instead of a pressed cone" or whatever.
They would cook something up for you and send you a few dozen sample cones, then rinse repeat until you got what you wanted.
I was puzzled as to how these guys had any objective sense of what they were doing. Was it all just witchcraft? Or did they have measurement equipment?
The chief engineer was this guy named George Pope. I asked him, "George do you have some speaker-driving fixture around here, that you stick all these cones on so you can run your own tests?"
"Nope. Here, let me show you how I evaluate a speaker cone." He walks us into the bathroom.
He says "This bathroom is super reverberant, it echoes for half a second when you snap your fingers. I just walk in here with a speaker cone and I flick the edge of the cone. A hard pressed paper cone makes a sharp crisp cracking noise. A soft felted cone makes a dull cardboardy sound. A poly cone makes a plastic-y sound. A metal cone makes a metallic sound. I don't measure anything. I just snap the edge of the cone in the men's room and listen to the echo. That tells me exactly what the cone is going to sound like."
Dang was that ever an epiphany for me. Because George flicking the cone is an impulse; and what you hear with your ears is an FFT.
That cone material has an intrinsic signature sound that is independent of the shape or size of the cone or whatever else it's attached to. It has its own signature sound.
From that point forward it was obvious to me that pressed cones have a certain sound, felted cones have a certain sound, dense vs. not dense, plastic vs paper vs metal. Same with tweeter domes. Same with compression driver diaphragms.
Same with box materials. Particleboard does not sound like plywood which does not sound like injection molded plastic which does not sound like marble which does not sound like machined aluminum.
Today in the age of DSP when you can EQ anything "ruler flat" with the touch of a button, you still cannot make a paper cone tweeter sound like a soft dome tweeter, which does not sound like titanium which does not sound like aluminum which does not sound like beryllium.
I'm pretty sure you can measure some of those differences which you so easily hear, but they do not easily or readily show up in the measurements. They may not show up at all in the frequency response curves. I suspect there are characteristics which you would be hard pressed to define purely in terms of measurements.
So I can't help but wonder if Alnico likewise has its own signature sound. How can it not? It has very peculiar hysteresis / saturation curves etc.
Ditto ferrite.
Ditto neodymium.
Materials have their own signature sound, no matter how they are used. Right now we're talking about magnetics.
How about BJTs vs FETs vs JFETS vs class A vs AB vs various class D implementations? Or triodes or pentodes?
Just like EQ can't make a metal dome sound like a soft dome, maybe 40 dB of negative feedback can't make a BJT sound like a MOSFET either. You can get close but there is always a "grain" in the sound that the trained ear can detect.
The "objective measurement people" are always skeptical about this, but the subjective people consistently maintain their preferences over long periods of time.
Maybe Alnico has some secret sauce that ferrite doesn't have
Last edited:
Hey just curious in your hybrid you put half in a sealed enclosure and the other half sticking out but it's technically a ported box if you think about the opening of the woofer when it's exposed so the air would leave the box going thru the opening of the exposed woofer could you have put felt over the opening to slow the air velocity of the air rushing thru the openings would it make a difference
Hi Lynn,
I am little surprised about your mentioning TAD drivers. In several of your previous posts you have attributed the performance of the Altec/GPA to the under hung voice coil constructions.
Although the TAD drivers have, likewise, under hung under hung voice coil construction according to TAD literature, cf. , post #72 in the link below, there is an interesting analysis disputing the claim, cf. https://www.audioheritage.org/vbulletin/showthread.php?20088-Explain-Tad-woofers-Help.
Perhaps, there are other attributes, contributing to your preference, e.g., efficiency, Fs, Qt, cone weight, mechanical damping, mentioned by marco-gea and discussed with dr. Geddes, that (further) distinguish these drivers from the current offerings?
Please do not misunderstand, I am not disputing your preference. I just question attribution of the preference to a single parameter.
Kindest regards,
M
I am little surprised about your mentioning TAD drivers. In several of your previous posts you have attributed the performance of the Altec/GPA to the under hung voice coil constructions.
Although the TAD drivers have, likewise, under hung under hung voice coil construction according to TAD literature, cf. , post #72 in the link below, there is an interesting analysis disputing the claim, cf. https://www.audioheritage.org/vbulletin/showthread.php?20088-Explain-Tad-woofers-Help.
Perhaps, there are other attributes, contributing to your preference, e.g., efficiency, Fs, Qt, cone weight, mechanical damping, mentioned by marco-gea and discussed with dr. Geddes, that (further) distinguish these drivers from the current offerings?
Please do not misunderstand, I am not disputing your preference. I just question attribution of the preference to a single parameter.
Kindest regards,
M
100% agree all materials have a sound, with no exceptions. An experienced listener can instantly hear them ... it only takes me a few seconds to assess a speaker or an amplifier, which makes going to hifi shows less than fun.
EQ does some things, but it does not remove the sound of the material itself ... it just scrapes off some of the annoyance. This is not so surprising when we consider that EQ does not actually correct multiple diaphragm modes that radiate off in different directions (a disruption of the polar pattern), it just modifies the power going into the voice coil. It's a one-dimensional correction for a nonlinear driver that radiates into 3-dimensional space ... that also has local reflections in the 3 to 10 mSec range (front wall, side wall, and floor).
So EQ, or servo feedback, work best at the bottom of the frequency range, where behavior is simpler (and models more accurate). I'm a big fan of Rythmik servo subwoofers, which use feedback to both define the frequency response (independent of cabinet volume) and reduce distortion. This makes a lot of sense below 100 Hz. Above 400 Hz, I don't find response shaping (via EQ) that convincing, no matter what the mike says.
I suspect (without proof) that resonances commingle with localized regions of high distortion, as well as disturbances of polar pattern. These problem areas are caused by the cone departing from the simple model and acquiring more complex high-order behavior, which affects all domains at once ... frequency response peaks, energy storage in the time domain, fluctuations in radiation pattern, IM distortion, all at the same time. All of our measurement techniques are nothing more than indirect measures of physical behavior of cones, diaphragms, enclosure wall flexing, internal modes, and edge diffraction. These things are physical and real; the measurements are oblique and indirect, and are primarily useful as diagnostic tools to find out what the physical problems are.
What we hear are the physics of the loudspeaker, not the frequency or time response. And human ears are very very good at telling what things are made of, since evolution adapted us not just for a high-risk environment but also musical enjoyment over at least 30,000 years of history.
EQ does some things, but it does not remove the sound of the material itself ... it just scrapes off some of the annoyance. This is not so surprising when we consider that EQ does not actually correct multiple diaphragm modes that radiate off in different directions (a disruption of the polar pattern), it just modifies the power going into the voice coil. It's a one-dimensional correction for a nonlinear driver that radiates into 3-dimensional space ... that also has local reflections in the 3 to 10 mSec range (front wall, side wall, and floor).
So EQ, or servo feedback, work best at the bottom of the frequency range, where behavior is simpler (and models more accurate). I'm a big fan of Rythmik servo subwoofers, which use feedback to both define the frequency response (independent of cabinet volume) and reduce distortion. This makes a lot of sense below 100 Hz. Above 400 Hz, I don't find response shaping (via EQ) that convincing, no matter what the mike says.
I suspect (without proof) that resonances commingle with localized regions of high distortion, as well as disturbances of polar pattern. These problem areas are caused by the cone departing from the simple model and acquiring more complex high-order behavior, which affects all domains at once ... frequency response peaks, energy storage in the time domain, fluctuations in radiation pattern, IM distortion, all at the same time. All of our measurement techniques are nothing more than indirect measures of physical behavior of cones, diaphragms, enclosure wall flexing, internal modes, and edge diffraction. These things are physical and real; the measurements are oblique and indirect, and are primarily useful as diagnostic tools to find out what the physical problems are.
What we hear are the physics of the loudspeaker, not the frequency or time response. And human ears are very very good at telling what things are made of, since evolution adapted us not just for a high-risk environment but also musical enjoyment over at least 30,000 years of history.
Last edited:
Good point. A resistive-vent enclosure is a totally different animal than a vented enclosure. The alignment is quite different, more like a closed box with a somewhat higher cutoff. The main virtue of a resistive-vent enclosure is insensitivity to shifts in driver Qt and amplifier output impedance. By contrast, a vented enclosure, being a 4th-order highpass filter, is fairly sensitive to Qt shifts in the driver. 4th-order filters of any kind are pretty sensitive to detuning, while 2nd-order filters are quite relaxed.
Hint: If your power amp is a zero-feedback design with a Zout of 1.5 ohm or higher, maybe 4th-order crossovers are not for you.
Last edited:
@dawa I only connect the MiniDSP to my computer for testing, and I use the USB input for that.
MiniDSP 2x4HD has USB, analog and fiber optic inputs.
For music listening I use a WiiM streamer https://amzn.to/40b5lqL which costs $90. I stream TIDAL (lossless) from my smart phone to the WiiM unit via Apple Airplay on WiFi. The WiiM connects to the MiniDSP via fiber optic cable.
I don't use a preamp. Most preamps are just unnecessary middle management that reduces signal quality and costs money.
I control volume via the phone and use the MiniDSP remote to set master volume.
I believe the MiniDSP goes up to 96KHz bandwidth and the WiiM supports that.
One issue: The MiniDSP has a high enough noise floor that on a typical amp with 26dB gain + Bitches Brew tweeters that put out 106dB for one watt, you can hear the MiniDSP's noise when no music is playing. At first I thought it was from the amps, but it is the 2x4HD. I had to insert 12dB attenuators between the DSP and the amp and that fixed the problem.
In a different system I have an Adcom GFA 2535 and I just turn down the input level on the back panel and it solves the noise problem.
MiniDSP 2x4HD has USB, analog and fiber optic inputs.
For music listening I use a WiiM streamer https://amzn.to/40b5lqL which costs $90. I stream TIDAL (lossless) from my smart phone to the WiiM unit via Apple Airplay on WiFi. The WiiM connects to the MiniDSP via fiber optic cable.
I don't use a preamp. Most preamps are just unnecessary middle management that reduces signal quality and costs money.
I control volume via the phone and use the MiniDSP remote to set master volume.
I believe the MiniDSP goes up to 96KHz bandwidth and the WiiM supports that.
One issue: The MiniDSP has a high enough noise floor that on a typical amp with 26dB gain + Bitches Brew tweeters that put out 106dB for one watt, you can hear the MiniDSP's noise when no music is playing. At first I thought it was from the amps, but it is the 2x4HD. I had to insert 12dB attenuators between the DSP and the amp and that fixed the problem.
In a different system I have an Adcom GFA 2535 and I just turn down the input level on the back panel and it solves the noise problem.
Hi @perrymarshall. I've been thinking more about the Bitches Brew and have a couple questions:
1. My Orion dipoles can rock back and forth at spirited volume levels, especially when on a plush carpet. Have you noticed anything similar with the Bitches Brew? I assume the beefy mass prevents this issue. Later versions of the Orion had a force cancelling woofer configuration to solve this problem. I've also had issues on wooden floors in older homes where the floor will resonate too much. Force cancelling fixes this as well. That said, I like the clean look of the Bitches Brew as it is.
2. Have you looked in to the Radian 15" 5215B Beryllium coaxial driver? Surprisingly, it's about the same price as the B&C 15CXN88 titanium coaxial. B&C has more data showing good polors, backed up by your measurements. There doesn't seem to be much info out there about the Radian except for what's on the spec sheet.
3. I notice you have a foil inductor in the rear photo. Is that for the rear tweeter notch filter? Can you comment on your passive part selection? I'm curious if you have rules of thumb around this or "go to" brands you use for the the passive components.
Thanks!
1. My Orion dipoles can rock back and forth at spirited volume levels, especially when on a plush carpet. Have you noticed anything similar with the Bitches Brew? I assume the beefy mass prevents this issue. Later versions of the Orion had a force cancelling woofer configuration to solve this problem. I've also had issues on wooden floors in older homes where the floor will resonate too much. Force cancelling fixes this as well. That said, I like the clean look of the Bitches Brew as it is.
2. Have you looked in to the Radian 15" 5215B Beryllium coaxial driver? Surprisingly, it's about the same price as the B&C 15CXN88 titanium coaxial. B&C has more data showing good polors, backed up by your measurements. There doesn't seem to be much info out there about the Radian except for what's on the spec sheet.
3. I notice you have a foil inductor in the rear photo. Is that for the rear tweeter notch filter? Can you comment on your passive part selection? I'm curious if you have rules of thumb around this or "go to" brands you use for the the passive components.
Thanks!
@Yourmando 1) I've never had a problem with this. Valid concern though. The BB's weigh about 100 pounds.
2) Radian looks attractive, it has one drawback which is it has a narrowing radiation pattern above 10KHz which the B&C does not. That's why I chose the B&C. I do like Beryllium and my Birch Dipoles featured in AudioXpress now have Beryllium diaphragms in their Radian 5208 coaxes. A noticeable improvement.
3) Foil conductor is for the series crossover between the subwoofers and the 15" bass mid, 110Hz. See schematic. It's a 6mH air core 0.6 ohm DCR fat inductor. Capacitors are a mix of high quality electrolytics and large Dayton film capacitor. I'm not terribly superstitious about capacitors, I do like film types better than electrolytics. I prefer air core inductors and it's important to have low DCR here, because higher DCR allows more subwoofer bass into the midrange driver and increases mid excursion. Since the low bass is boosted +15dB from the DSP, it's important that the midrange driver is shielded from that.
FYI if I feed 18Hz into the system at ~30 volts / 200+ watts, the SB subs are flapping hard (however not bottoming) and the B&C 15CXN88 shows maybe 3mm of excursion peak-peak. The mids have about 1/10th the excursion of the subs which is as it should be.
2) Radian looks attractive, it has one drawback which is it has a narrowing radiation pattern above 10KHz which the B&C does not. That's why I chose the B&C. I do like Beryllium and my Birch Dipoles featured in AudioXpress now have Beryllium diaphragms in their Radian 5208 coaxes. A noticeable improvement.
3) Foil conductor is for the series crossover between the subwoofers and the 15" bass mid, 110Hz. See schematic. It's a 6mH air core 0.6 ohm DCR fat inductor. Capacitors are a mix of high quality electrolytics and large Dayton film capacitor. I'm not terribly superstitious about capacitors, I do like film types better than electrolytics. I prefer air core inductors and it's important to have low DCR here, because higher DCR allows more subwoofer bass into the midrange driver and increases mid excursion. Since the low bass is boosted +15dB from the DSP, it's important that the midrange driver is shielded from that.
FYI if I feed 18Hz into the system at ~30 volts / 200+ watts, the SB subs are flapping hard (however not bottoming) and the B&C 15CXN88 shows maybe 3mm of excursion peak-peak. The mids have about 1/10th the excursion of the subs which is as it should be.
Last edited:
Hi. I am newly baffled. I am a new baffler? I have a few general questions for the general public of opinion. I would like to share my favorite speaker and why I believe it so, and attempt to mimic it in a smaller footprint. I was the proud owner of a pair of Frasier model 11's for 20 years until a guy offered me enough $$
I couldn't say no. They had 4, 4" mids 2 Piezo tweeters, and a 15 and 12 " woofers. the Side walls of the entire cabinets created the vent. The soundstage was incredible. Snare drums cracked like no other. The mids were so impressive. The piezo had a cloth lining that stopped the ring. They were about 5 feet high and 3 feet wide. Lots of frontal surface made the difference I think.
I made myself feel better after the loss as i decided to build a set. do you think there would be a way to mimic the big baffle by mildly creating a concave surface on the front baffle? Directionality would be reduced by dipole mids and tweets. I am going to box the woofer. Thoughts?
I couldn't say no. They had 4, 4" mids 2 Piezo tweeters, and a 15 and 12 " woofers. the Side walls of the entire cabinets created the vent. The soundstage was incredible. Snare drums cracked like no other. The mids were so impressive. The piezo had a cloth lining that stopped the ring. They were about 5 feet high and 3 feet wide. Lots of frontal surface made the difference I think.
I made myself feel better after the loss as i decided to build a set. do you think there would be a way to mimic the big baffle by mildly creating a concave surface on the front baffle? Directionality would be reduced by dipole mids and tweets. I am going to box the woofer. Thoughts?