Beyond the Ariel

[Salas]

Originally posted by Lynn Olson


I'm planning on the midbass and bass driver sharing a steeper 12 dB/octave crossover set to about 1 kHz to prevent mid colorations from sneaking in. The cascade of 6 dB/oct lowpasses for the midbass and bass will control the degree of overlap between the drivers, and allow response shaping without massive EQ from the power amps.

I suspect mildly dissimilar widerange and midbass drivers might be a good thing, actually - remember, all of these drivers are working in the piston band, with intrinsically flat response, so what we're hearing are IM distortion signatures and unusual things like "magnet sound".

As for the bass driver, either a 15 or 18 would work fine. This depends on sonics as far as I can tell. High Q is a must, of course, along with good efficiency to complement the 97 ~ 99 dB/metre drivers upstairs.

The 18inch is mainly used to save some excursion for +3 to +4dB on top of a 15 inch territory output. It is generally heavy, slow, and beams from 100-150Hz and up. Generally its parameters don't give lower enough bass than a 15''. Main reason of use is max SPL. We are talking outdoors reinforcement levels now. There, room gain is absent and ground image is the only aid. Thus you get em in twin 18'' side by side mounted format in wide and low 200lt and over reflex or bass bin boxes. Processors cut like hell under 35Hz to save excursion limit, and system phase goes south.
I have put 18 and 15 together in shared box volume, only there they time better. I would choose no more than 15'' in the speaker under discussion (no enclosed volume to use the shared timing trick anyway) and if I needed more stuff for EQ and raw level, I would just double up.

[planet10]

Originally posted by planet10
The ladder structure at the back doesn't feel right, there must be a more elegant way of achieving the same ends.

Here are some pics of other efforts... these thou, just brace the baffle, but could be adapted with the addition of a magnet clamp amd a tripod structure on the front/baffle end with driver mounting rings... (somewhere on this forum is a Thorsten L supravox baffle with a metal L for driver mounting and a floating plexi baffle)

The big new Jamo flagship:



MDF -- not the ideal material...



and a rendering that preceded the above....



dave

[LineSource]

Lynn, have you considered a 2-structure baffle to separate the dipole woofers?

Dipole woofers will generate baffle vibration that can audibly modulate the higher frequencies unless special attention is paid to the baffle design. Counter balanced W-baffles are one good solution for deep bass from multiple dipole woofers. My earlier post illustrated how sand cavities can be built into the edge round-overs to maintain forward facing dipole woofers. I favor forward facing dipole woofers, but always physically separate the bass structure from the mid-tweeter structure.

[Graham Maynard]

http://www.jamo.com/Default.aspx?ID=...roductID=17862

[johninCR]

"MDF -- not the ideal material..."

I'm starting to think that for OB's it may be pretty close.

Dave, I like the spine approach, especially if you connect it to the magnets. Add a couple of sacks of steel shot to the base and you're in business.

I agree with linesource regarding firing the woofers on axis with physical separation of the woofer section. Also plan on mass even in the upper section if you want a vibration free baffle. It's surprising the forces at play even in the null of the woofer section, and I'm not even talking about high output like Lynn's goal. They only net to zero there, not actually cancel.

It's great to see all you box guys finally coming around. The next step is to get you away from the flat baffles and H's, especially in the LF section. That's where the opportunity lies for design advancement. Anyone can XO and EQ a multiway flat if I can. As a group I'm sure we can figure out how to optimize my dual pathway approach in the LF section, or are you guys going to leave me on my own for another year. If I go it alone, I may become averse to sharing and go commercial.

[Lynn Olson]

Originally posted by LineSource
Lynn, have you considered a 2-structure baffle to separate the dipole woofers?

Dipole woofers will generate baffle vibration that can audibly modulate the higher frequencies unless special attention is paid to the baffle design. Counter balanced W-baffles are one good solution for deep bass from multiple dipole woofers. My earlier post illustrated how sand cavities can be built into the edge round-overs to maintain forward facing dipole woofers. I favor forward facing dipole woofers, but always physically separate the bass structure from the mid-tweeter structure.

The catch here is that I'd like to run the wideband driver pretty much full range, without a 200~300 Hz highpass. If I was going to use a highpass filter - and make an honest midrange of it - the obvious candidate is the 18Sound 8NMB420. Note the response curve and power handling. For a classic 3-way speaker, this is an obvious design choice, and gives more freedom to select the tweeter since a higher tweeter crossover can be selected.

Since the 12NDA520 is the target driver, and I plan to let it "free run", it will be generating some cone-reaction vibration. Not as much as the drivers below it, of course, and that's a good argument for physically isolating the wideband driver from the midbass and bass drivers.

Planet10, thanks for the posts at #448 and #452. Good ideas there, and the 3D illustration is a slick visualization tool for the driver-support framework.

It's interesting - a conventional enclosure is like a monocoque, or exoskeleton, with the outer shell bearing the structural weight but also unfortunately acting like a large-area radiator for all internal vibrations, whether reaction products from the driver, or standing waves from internal box modes. The large area is much of the reason minimonitors have less cabinet coloration than even the most advanced construction big speakers - there's just less area to radiate, thus less coloration.

By contrast, the driver-support structures we're seeing here are external, but are specifically designed to have low radiating area combined with stiffness and rigidity. The trick is to have the large-area baffle not flop around and resonate from the back pressures, which although low, are certainly present in close proximity to the drivers.

Originally posted by johninCR
"MDF -- not the ideal material..."

I'm starting to think that for OB's it may be pretty close.

Dave, I like the spine approach, especially if you connect it to the magnets. Add a couple of sacks of steel shot to the base and you're in business.

I agree with linesource regarding firing the woofers on axis with physical separation of the woofer section. Also plan on mass even in the upper section if you want a vibration free baffle. It's surprising the forces at play even in the null of the woofer section, and I'm not even talking about high output like Lynn's goal. They only net to zero there, not actually cancel.

It's great to see all you box guys finally coming around. The next step is to get you away from the flat baffles and H's, especially in the LF section. That's where the opportunity lies for design advancement. Anyone can XO and EQ a multiway flat if I can. As a group I'm sure we can figure out how to optimize my dual pathway approach in the LF section, or are you guys going to leave me on my own for another year. If I go it alone, I may become averse to sharing and go commercial.

I was thinking again about user-adjustable hinged baffles - with the simple addition of airtight tape on the back side, a piano hinge would indeed work to create a variable-angle side wing. Any kind of wing is going to create a short horn. In this application the de facto "horn" is working far below its passband, and moreover is looking directly into a room corner!

As you've been mentioning, JohninCR, this rotates the null of the dipole polar pattern, which in turn changes the interaction with the room, which is nowhere close to flat at these frequencies anyway. Cheaper than EQ, and certainly saves on the amplifier power.

The sheer strangeness of combining a TQWT in acoustic parallel with a leaky open-ended box has an appeal, although how it could be modelled as a complete in-room system escapes me. The people doing the heavy-duty competer modelling appear to have their hands full with simple dipole baffles interacting with a room. The simplest possible model of the TQWT+lossy-box would seem to shift from a monopole to cardioid to a dipole as the frequency increased, but that's just a wild guess. It's going to interact with the room in a very different way than conventional monopoles or even large-panel dipoles.

As JohninCR has mentioned earlier, measurements don't correlate well with subjective impressions at these frequencies, so the merit of going to the trouble of modelling all that is open to question. And I thought TL's and Karlson enclosures were controversial!

[johninCR]

Lynn,

The problem with hinges is bracing. This is without question the most common flaw I see in virtually every folded open alignment. The result is panel vibrations that would never be acceptable in any decent box speaker. Do these people really not put their hands on the construction to feel the significant vibrations. Apparently the common thought process is that since it's opened, there is no pressure, however, boxes don't vibration from a change in absolute pressure in the box. It's the pressure waves striking them. If you constrain that rear wave to a significant extent those rear edges "flap in the wind". Bracing may be even more important than a box, because at least a box has the structural support of the rear corners.

[Lynn Olson]

Just wanted to throw out some information on vibration control, looked at from two different angles:

First, the "mainstream" approach of using every known method to get it as low as possible. Thanks to "jzagaza" at http://www.diyaudio.com/forums/showt...52#post1165952, we have some very interesting links to Fujitsu's advanced audio research project at:

http://www.timedomain.co.jp/tech/tech_e.html

http://www.timedomain.co.jp/tech/the...eoryA4_eng.pdf

And a rare Stereophile review that actually describes the real-world results of this project - compared to something completely different, an Avante-Garde all-horn system!

http://stereophile.com/standloudspeakers/107fuj/

By the way, the Fujitsu Ten Eclipse TD712z speaker is about as close to an ideal monopole - in terms of extremely low diffraction and fully decoupled vibration control - as any speaker I've ever seen. This seems to represent the limiting case for the "mainstream" method of rigorous isolation and mechanical sinking of vibration from the driver. The physically small size of the single-driver speaker also limits unwanted emissions from the cabinet, something that is quite troublesome in big audiophile (and dipole!) speakers.

And for the other view - instead trying to completely get rid of vibration, over at Mother of Tone, they're aiming for "consonant" vibration, using selected woods and varnishes:

http://www.mother-of-tone.com/mother.htm

http://www.mother-of-tone.com/lacquer.htm

This isn't as ridiculous as it sounds. Thom Mackris and I auditioned identical compressed-carbon (not carbon-fibre) record mats with, and without, a common woodworking varnish in a direct A-B test. The difference was equivalant to the difference between a $500 and $5000 tone-arm, or a $200 and a $2000 phono cartridge. More musical, as you'd expect, but also much more lifelike and a remarkable absence of "mechanical" or metallic coloration. Instead of adding a "musical" coloration, it removed quite a bit of harsh and grainy "mechanical" coloration, and revealed much more of the musical dynamics, tone colors, and timbre. The impression of space was also quite a lot better, remarkably so, and it didn't sound like a coloration to me, more like the removal of a number of resonances.

Are the "absolute" and "consonant" approaches in conflict? Not as much as it appears. Yes, it's good to get coloration as low as possible, but there's always going to be a residue. That residue might as well be musically consonant, so it can subjectively disappear. The Mother of Tone approach using solid woods and selected lacquers (not C37, by the way, but conventional wood lacquers) seems like a direct and cost-effective way to build agreeable-sounding speaker systems using commonly available materials.

[ScottG]

While the 12NDA520 is a fantastic driver.. I'm not so sure this is the driver that should be run fullrange.. Dispersion becomes poor about 1.4 kHz, and it looks like at 1.8 kHz you start into time domain problems.

Now of course you *might* be able to modify the driver to an extent to clear things up (in time).. but that still won't do anything for the dispersion pattern.

It could be that the Tone Tubby is FAR better with time at the top of its passband.. don't know. I also don't know about its horizontal dispersion (which, including price, is why I didn't mention it to Paul in his thread).

One possibility to improve in-room power response while even further limiting dispersion, (creating a more linear off-axis response), and near time room reflections would be a hypercardoid based on a free air bipole with a pair of 12" drivers (mounted back to back). It would however need a very similar dispersion pattern at higher freq.s - which could be accommodated via those 18 sound horns. (in a paired bipole config.). Note that such a config. for the free air bipole would significantly reduce frame energy build up, (provided the two drivers were connected physically via a high transmission material), and mildly reduce 2nd order distortion. More importantly though, it should reduce time problems at the top of the driver's passband (..when compared to virtually any other mounting scheme). Additionally, you could always vary the rear in-phase output for most of the passband with a simple acoustic trap.

(Note that such a bipole will NOT have the side combing problems of traditional bipoles).

Well, it's an idea.

[berm]

I’m curious about the worth of this approach, but I’ve been intrigued by a woofer design in Speaker Builder
2/93, “The Flexible Dipole Woofer”, by Michael Allen. It’s part of a 3-way dipole, and covers the range between 40 and 135 (?) Hz. The whole system he claims to be + or - 2dB, 40 to 20kHz, with tight, tuneful bass.
He constructed the woofer on a frame 33” by 22”. On that he glued a very thin sheet of mylar, then on that glued a 29” by 18” piece of polyurethane foam (soft foam, I’ve assumed), then another piece of mylar glued over that to form a sandwich. He then takes a 10” woofer, cuts off most of the basket save for two legs, removes the cone and dustcap, leaving it as a motor. He mounts this on a double bracket on one side of the frame, and attaches it to the foam sandwich, using the neck of a 2 liter pop bottle inserted into the 1” voice coil, then glued to the mylar. He goes on to write that he was surprised to find that Museatex has a patent on a similar device, but he sees his as a “soft piston” and the Museatex as a “bending wave” transducer. I believe the Museatex design is the Melior One, a single-driver point souce which I think moves a mylar film without the foam sandwich feature.
This is one of those projects that I never got around to trying. This design should have several applications, and it seems like this particular thread was the best place to drop this in as one way among many to get decent dipole bass. I’ve been amazed at how occasionally some ideas seem to take fire, and the recent threads - this one, EnABL/Mamboni, the fourteen inch handmade cone, the Walsh thread - are enlightening and fun to read, to say the least. I should have dropped this in earlier, as the discussion seems to be going in a different direction, but I thought it still might prove useful to someone - if it does merit consideration.
Keep all that gray matter churning. Great read on this thread.