A “Good Enough” OB/Sealed Hybrid for “Everyman”?

I know this is not a new topic but perhaps my PoV about it is different enough for this post to be of interest. Pls bear with me.

Ken Kantor’s NHT 3.3 speaker took advantage of the bass reinforcement effects of room boundaries with a low Q sealed enclosure to provide predictable deep bass extension. This large late ‘80s 4-way combined a 12” side-firing woofer positioned at the intersection of floor and front wall with an ear-height assembly of lower mid, high mid & tweeter. The 3 smaller drivers were at the top of a front baffle angled 21 degrees toward the listening area; in mirror image pairs, the woofers for a stereo pair faced each other. They had an unusual shape and size: 42" H by 7" W by 31" D. No typo -- just 7” wide but 31” deep! The angled baffle made it just wide enough to fit the 6.5” low mid driver. The narrow deep shape was deliberate. The upper frequency drivers ended up >2.5’ away from the front wall, which reduced diffraction and increased the ratio of direct-to-reflected sound. These speakers were lauded for deep articulate bass even at high levels, good imaging and neutral balance. https://www.stereophile.com/floorloudspeakers/1293nht/index.html

nht-33-audiophile-loudspeakers.jpg


I ran the smaller NHT 2.9 in my main system for some years. The woofer was a 10” instead of the 12” in the 3.3, and the cabinet was scaled down, but the other drivers were identical to the 3.3. It was not the last word in immediacy or transparency, but those aspects were decent, and the overall tonal balance when placed close to the short wall of my room (13’ x 30’) was excellent. Bass was deep and powerful if a touch heavy on some recordings, and the speakers were tonally neutral, not favoring any particular type of music. They were quite enjoyable.

The Linkwitz Orion (3.3) I built in 2014 outperformed the NHT 2.9 in every way, but this was to be expected. The Orion was active, EQ'd electronically, fully dipole, and employed drivers that cost upwards of CA$3,000 -- many times that of the components in the NHT. It was also driven with 50Wx8 rather than a 100W/ch stereo amp & needed to be further out away from the wall to do their magic.
orions.jpg


I’ve been experimenting with open baffle speakers ever since. My latest LX521-inspired speakers are likely as good as I’ll ever want. They are amazing performers that evoke “wow” from just about anyone who listens to them. But much as I love the imaging, depth & openness of open baffle speakers, their complexity, cost and positioning requirements are not for everyone. My LX521-inspired build, for example, uses drivers of ~CA$2,500 net, 8-conductors of cable to each speaker, an 8-ch Hypex NC252 amp connected with 8 signal cables to a minidsp 4x10 HD. In my audio-dedicated acoustically treated 22Wx19’D studio, they sound best nearly 5 feet from the front wall and >6’ from the side walls. Such positioning plus >CA$7,000 in components (before any source or control components) aren't feasible for a lot of folks. (Never mind the US$24,000 for a factory-built LX521 system from Linkwitz.store)

lx521.jpg

A couple of close friends have asked if I could create cheaper, less complex speakers that can be positioned up against the wall yet retain much of the magic of my open baffles. It needs to have rich, clean bass, balanced FR, nice transient response, spacious imaging, be easy to drive, be reasonably tolerant of poorer recordings, and sound good from modest to fairly loud levels. It’s a call for an inexpensive, full bandwidth loudspeaker that images well, sounds big & clean, does nothing badly and isn’t very demanding. An entry level high end speaker that some would be happy to stick with for the long run.

I’ve been pondering this a lot. I know the shape, dimensions and acoustical treatment of the studio is a huge factor of my OB system performance. But getting most people to understand and act upon the need for acoustic balancing of their rooms is difficult. So focusing first on suitable speakers, one train of thought led me to a hybrid of open baffle & the NHT boundary-reinforced sealed woofer design.

A 3-way tower speaker consisting of…

  1. NHT floor/wall-adjacent style bass driver in low-Q sealed enclosure
  2. Minimal-baffle dipole for mid & tweeter, rotatable for “aiming”
  3. Deep tallish cabinet to put the dipole upper section at ear height far enough from the wall.
  4. Crossover points & slopes chosen for smooth transition from omni in the bass to narrower at the top. Also avoid floor-bounce null?
  5. Dayton Audio KABD-4100 4x100W + DSP + Bluetooth 5.0 modded as plate amp -- or passive crossover after development with DSP & active amplification
  6. The target max cost for all components is CA$1500/pair.
 
Last edited:
  • Like
Reactions: 3 users
To further explore the above criteria:

  1. NHT floor/wall-adjacent bass driver in low-Q sealed enclosure -- a fairly reliable way to obtain articulate sub-40Hz bass. Not as good as OB, imo, but sealed bass generally sound better than most ported ones to me, and have better overload protection from low frequency grunge in vinyl sources. A system Q of 0.5 is ideal (I believe) for quarter-hemispherical radiation, but if the box is made a bit too small and higher Q results, then positioning a bit further from the front wall can counter any bass hump -- but many prefer a “fat” bass over a more linear presentation anyway. The question of crossover frequency and slope for the inside-facing woofer comes up. The NHTs had 2nd order crossovers at 100Hz & 320Hz, and 3rd order at 3.5 kHz. The 100 Hz xover point may have been chosen to prevent directional confusion as the woofer is mounted sideways. However, a 100 Hz xover point is challenging for an open baffle driver that might also need to go as high as 1500 or 2000 Hz. 100 Hz almost demands at least an 8” mid driver, a size that limits how high it can go. Perhaps a 6.5” mid should make it to 1500 Hz without much beaming. Also, would a front-facing woofer get enough reinforcement for good bass extension? If so, crossing to a 6.5” mid at 300 Hz or higher is probably preferable.

The Peerless 830668 10” SLS, used in John K’s NaO Note II, seems a good inexpensive woofer choice: A sealed 1.6’ enclosure would put F3 at around 45 Hz. (Key parameters: 32 Hz F, 0.55 Qts, 62 liters Vas) Judicious stuffing & placement against the wall should bring F3 down below 40 Hz. EQ with DSP would make this easy. The stamped frame is less than ideal but pressure-bracing against the back of the magnet should reduce effects of frame vibration. Unlike so many Peerless drivers, this one is still available from Solen & Parts Express in <1000-lot quantities.​
A box of ¾” thick panels with outer dimensions 24”D x 30”H x 5.8”W is 1.6 cubic feet. Increasing the width of the front panel would make it look more balanced with the open baffle module on top, so the height could be dropped to 28” and width increased to 8”. Any extra volume won’t hurt.​

  1. Minimal-baffle dipole for mid & tweeter, rotatable for “aiming”. There are so many examples of minimal baffle dipoles, and much discussion of the benefits (and drawbacks) so no need to elaborate. The main question is how much distance between wall and back of drivers is needed to get at least some of the dipole benefits? I aim to determine that experimentally, but if others can chime in with experiences or simulations, that would be most helpful.
I have a handful of new aluminum cone 6.5” SEAS L18RN drivers. They were custom made for a Canadian startup that didn’t during the covid pandemic. Optimized for low bass but with decent measured FR to 1.5 kHz. Well made, typical SEAS quality. This is my choice for the midrange ATM.​
Dipole tweeters like the GRS magnetic planars and small AMTs were considered, but none are able to extend down low enough. 1500 Hz is a tough call for most small tweeters. However, XRK971’s well-liked FAST / WAW Ref Monitor uses the Scanspeak 10F full-range 3” driver as a tweeter, which got me scouring my shelves for similar drivers. The 10F (used in my version of LX521) is a bit too pricey to meet my (semi-arbitrary) cost target, but a Peerless TC9FD18-08 is just 20% of the price, and fairly close in performance -- iirc, they did originate from the same developmental team. The magnet structure looks like it impedes the back radiation & the plastic frame is a little chintzy, but…​
The aiming part is based on the notion of angling the dipole to minimize side wall reflections. In some rooms this can be very beneficial, especially with a 3" full range that beams considerably more with increasing frequency than the typical 1" dome tweeter. A mid/tweeter dipole module with a bolt through its base to the top panel of the bass box would allow it to be turned as desired, then locked down.​
For my first try at the mid/tweeter module, I adopted the no-baffle scheme used in the Vistaton Nobbi open baffle speaker kit. I like the simplicity of the perpendicular brace, though the small flange of the TC9 driver makes it less than perfectly stable with just 2 screws. I will sort this out later if necessary. If dipole cancellation causes too much low end rolloff, I’ll try something closer to the LX521 baffle or larger. https://www.lautsprechershop.de/hifi/nobbi.htm
IMG_20230706_113040074.jpg
IMG_20230706_113102203.jpg

  1. Deep tall cabinet -- As mentioned earlier, this is where testing & observations with dipoles in many different rooms would be helpful. How far does the front wall have to be for clear dipole benefits? I suspect that a 31” deep speaker like the NHT 3.3 is tough for domestic acceptance. 24” is more sellable, hopefully.
Anyone with dipole speakers can try moving them closer to the front wall and gauge subjectively. Hopefully a few DIYaudio members will do so and chime in?​
  1. Crossover points & slopes chosen for smooth transition from omni in the bass to narrow at the top. Also avoid floor-bounce null? This is mostly a matter of juggling driver directivity, distortion I(and/or excursion limits), and filter slopes. The woofer/mid point would be ideal at 300~400 Hz to help counter early floor bounce cancellation, especially if using shallower 2nd order filters, I think. This also keeps the 6.5” driver from having to handle lots of EQ boost at lower frequencies. The mid/tweeter point seems less complex -- anywhere between 800 to 1500 Hz would likely work for constant directivity through the xover overlap with the current drivers. I expect the best frequency & slope can only be determined iteratively with measurements & listening.
  2. Dayton Audio KABD-4100 4x100W + DSP + Bluetooth 5.0 modded as plate amp -- or passive crossover after development with DSP & active amplification. This inexpensive product will probably raise eyebrows. It has been noted by other members here but seemingly never tried by anyone. It appears identical to the Wondom JAB5. The specs are adequate, but the claimed flexibility & features are amazing. It requires a 36VDC 10A power supply for full output, and such devices can be had for under CA$50 on Amazon, so I think it’s worth a try. https://solen.ca/en/products/dayton-audio-kabd-4100-4-x-100w-bluetooth-amp-board-with-dsp

The CA$1500 target for parts cost is low, but that’s part of what makes this project challenging. If your budget is unlimited, then no engineering is needed -- throw enough money at a technical problem and it can usually be solved. It’s when there are difficult targets within a limited budget that creative design, engineering & execution become essential and more fun.

PS -- It’s true that the Linkwitz LXmini might meet a lot of my criteria but it’s not likely to meet the critical domestic acceptance factor. My friends say there is no way they could sneak wired PVC pipes into their living rooms. :giggle:

PS2 -- Why just “good enough”? Because in many cases, it’s actually ideal. Context read:
https://positive-feedback.com/audio-discourse/can-a-system-be-too-good/

I welcome comments on any and all of the above.
 
Last edited:
  • Like
Reactions: 1 users
Perhaps the answer to this would be more forthcoming if you define what the benefits of a dipole are, since it isn't an ideal but rather a set of properties, each with pros and cons.
Thanks for responding, Allen. I really did think dipole characteristics are so thoroughly covered here that further discussion isn't necessary. I mean there's the 15 yr old 206 pagelong thread Ultimate Open Baffle Gallery -- among other things. :D Off the cuff, the benefits I hear & like with OB dipoles I've built & played with include great spatial imaging, sound that seem less localized to the drivers, greater insight into recording conditions, less sonic congestion... all of which are directly tied to the figure-8 directivity pattern & reduction of sidewall reflections. Then there's the whole set of mechanical benefits of zero back pressure on the driver cone, no resonances from an enclosed box, no backwave bouncing back through the driver cone or in the case of ported speakers, through the port.

But perhaps the lack of response to my post is more about the whole bass reinforcement at boundary junctions idea. I know I haven't seen the NHT-style designs in commercial offerings for quite a while. Is that notion now passe or proven to be too flawed in ways I don't know about? Or is it that 2' distance from the front wall is laughable for a dipole? I did have the back of the bass "box" in my dipoles as close as that sometimes, but that meant the drivers were >3' away.
 
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
I don't want to duplicate the well discussed issues on that thread. I see a middle ground range of behaviour where side reflections are reduced and rear reflections are increased. We know that late reflections are wanted, and for that they perhaps should be as many and varied as will reduce their distinctness. Therefore this is partially like a selective omnidirectional speaker.

There's not a more plain geometric method for fitting this radiation source to a room, it's strength appears to be to fit the early/late features of this type of speaker with the known properties of audibility. In one respect, reflections are best held off for at least 30ms, but no less than 10ms.
 
pretty much.

So either speakers pointed straight up, or straight down.

Such as a very large baffle called a Ceiling.

Or adding a basic diffusion delay line.
Aka Reverb effect to your mix.

You can also apply delay times to be frequency dependent.
Aka use the common 30ms spring reverb tank
 
Last edited:
Member
Joined 2018
Paid Member
Interesting project and I look forward to the developments. Sounds similar in theory to my recent build. Sealed bass and 'dipole' mid and hi but I probably don't get the figure 8 due to proximity to front and side walls .
IMG20230407173522.jpg
IMG20230402152953.jpg
No side firer mind you
 
  • Like
Reactions: 1 user
pretty much.

So either speakers pointed straight up, or straight down.

Such as a very large baffle called a Ceiling.

Or adding a basic diffusion delay line.
Aka Reverb effect to your mix.

You can also apply delay times to be frequency dependent.
Aka use the common 30ms spring reverb tank
I don't quite follow your shorthand. Could you pls elaborate?

AllenB, 30ms? Where does this # come from? A quick calc tells me sound travels about 1'/ms. That means the first reflection in my OB setup 5' from front wall reaches the listener under 10ms after the direct sound. The 30ms you cite would require the speaker to be 15' from the wall, which means 99% of OB users here are doing it wrong?

I can't cite any specific instances right now, but iirc, I've seen posts here (and elsewhere) of OB users lamenting adequate space behind the speaker but reporting they're still pleased with the end result. Pretty sure some were 2' or less from the wall.

I thought about placing a diffusion panel perhaps 2x2' or 3x3' on the wall directly behind the proposed OB array to compensate for the short distance.
 
  • Like
Reactions: 1 user
Interesting project and I look forward to the developments. Sounds similar in theory to my recent build. Sealed bass and 'dipole' mid and hi but I probably don't get the figure 8 due to proximity to front and side walls .

No side firer mind you
Ah, I noticed your thread -- https://www.diyaudio.com/community/threads/3-way-sealed-bass-ob-dipole-mid-hi.398072/ -- maybe should have just posted there! The woodwork of your post #24 looks excellent. Kudos!

Your xover also looks like it's coming along nicely. I still haven't broken into XSim or VtuixCAD yet -- need to do it soon.

Using 3 ch of a minidsp 2x4 hd with 1 speaker for crossover trials right now. And a random 2' box + woofer. A couple hours in, with LR4 at 350 Hz, LR2 at 1200 Hz, and a handful of PEQ adjustments, the FR looks decent, sound is not bad. Plenty of bass to below 30 Hz -- it got boosted by 8~10 dB below 50Hz when I moved the box from 2' out to the wall. Haven't turned it sideways yet. Overal tonal balance is OK, tho it needs a lot more work. Have to sit fairly close to tweeter axis for highs >6-8 kHz.

IMG_20230706_113604238.jpg
 
I thought about placing a diffusion panel perhaps 2x2' or 3x3' on the wall directly behind
You could also try a reflector near the rear of the driver to try to spread the back wave around. I expect it to lose effectiveness at lower frequencies (size dependent), but might still be useful. Similar to the old Sanders electrostatic beam splitter. I haven't tried it myself, but it would be quick and easy to experiment with wedges, balls, etc.
 
  • Like
Reactions: 1 users
30ms is where we can hear the two events separately without needing to think about it.
30ms.... That suggests audible echo, which is not exactly desirable, I'd think.

So if we define 10ms as the minimum delay between direct & reflected sound to prevent the blurring effect of diffraction, then >5' distance to the wall is necessary for any speaker with any significant backward output. This would include not only OB but many UK ported designs -- lots of BBC-originated speakers like Spendor, etc. Yet I've heard many OB systems (not just mine) that were positioned much less than 5' from the wall, and mostly been quite impressed with the clarity, imaging, etc -- no recollection of any blurring effects. Which makes me dubious about that 10ms.

Do you know the source of this 10ms & 30ms info?
 
OB is tricky to integrate properly with a room. Thats mainly why I don't use any OB designs. You need alot of wiggle room for placement, which I don't have. I do like the listening experience OB gives when properly set up and dialed in, but its a problem for my listening environment.
 
  • Like
Reactions: 1 user
OB is tricky to integrate properly with a room. Thats mainly why I don't use any OB designs. You need alot of wiggle room for placement, which I don't have. I do like the listening experience OB gives when properly set up and dialed in, but its a problem for my listening environment.
Well, that's why this is not a full OB. It's an attempt to get much of the benefits of OB with a close-wall footprint. It's not clear yet to me whether the 2' distance from wall to OB drivers is enough to provide that.

XRK's OB design is the only exception id explore, as it's proven in design and is small enough in footprint to successfully place in smaller rooms.
I presume you refer to his multiple woofer SLOB XSD? But that design is a full OB, and it has the same need for far-from-wall placement as all OBs. It might be a bit shorter, but its footprint is no smaller than many other OB dipoles -- including my 2 main ones, the Linkwitz Orions & LX521s.
 
After years of building speakers for other people , I decided on a project to my own tastes. No measurements, I just fine tuned my best guesses for the stereo speakers based on impressions and notes taken about these specific extended range 12 inches.
They started as flush mounted drivers off centred on wingless folded baffles.
But were a little lean in the lower male vocals, so got first a single deep wing on one side, then shallow wings all around, along with being sunk in the front about 3/8ths of an inch.
This got the sound from a smallish baffle acceptable with jazz or classical with a turntable without a sub, ditto for late-night apartment movie sound.
Settled on a simple H Frame push pull sub with an 80 hz crossover point for cinema sound.
Next up was finding the worst breakup point on the cones with a receiver's built in parametric equalizer, and approximating it with sort of a passive crossover, instead of a notch filter.
I settled on a shallow Q crossover point with widely spaced knees, centred about 8500 hz, and used a fairly decent tweeter about 5db less efficient than the main driver.
I do like treble, I just like it in lesser levels than most people. I didn't believe I'd like a rear firing tweeter, but tried a couple different ones anyways, and didn't keep them.
So I keep most of the cohesiveness of a single driver, get cleaner wider dispersion treble, fewer baffle reflections and more front wall reflections for spaciousness, and thanks to my neighbour allowing to do comparisons, more in my room bass with much less through the shared wall in theirs, even with some capable ported speakers at the same 80 db test tone playback.
Best of all, even playing music, with the sub less than a foot from my turntable at pretty exuberant levels, there's no feedback since the cartridge is right in the sub's null.
It took a few near hits and misses to get these this simple and still work to what I'm after.
 

Attachments

  • 20230707_140158.jpg
    20230707_140158.jpg
    731.7 KB · Views: 262
  • 20230707_135935.jpg
    20230707_135935.jpg
    531 KB · Views: 255
I didn't think 32" was all that far from the wall. For me it's a compromise between traditional 4 to 6 feet, and 2 or 3 feet as with a lot of speakers, or the home theater a few inches from the wall thing these days. It is pretty far out into the room compared to a lot of Paradigm and similar models I see around postings lately.
 
diyAudio Moderator
Joined 2008
Paid Member
With dipole designs the reflections are part of the effect and you craft them to be diverse and varied. It's a different scenario to a more typical speaker with specific reflections that are more distinct.
Do you know the source of this 10ms & 30ms info?
It's old knowledge. You suggest echos may be bad but once it becomes an echo it won't play a part in listening fatigue because we don't need to separate it.

https://www.diyaudio.com/community/threads/beyond-the-ariel.100392/post-4439329
 
  • Like
Reactions: 1 users