Enharmonic sound system

• Chart: Eighteen Sound's published 2.83 V response with their suggested EQ, redrawn by eye from the app note (±1 dB).
• Traced from the dimensioned side view in Eighteen Sound's 18H_v1 app note. The outer size, the 517 mm mouth and the panel corners are read from the drawing. Internal braces are left out.
• Path: the cone fires down into the throat under the driver panel, then up the back, across the top and out the front. The magnet sits in the sealed, damped rear chamber.
• 590 mm wide (23.2 in). 15 mm Baltic birch, M6 bolts in T-nuts, removable access panel for the driver. My weight estimate is 110–130 lb with driver.

• Chart: B&C's published TEF measurement with the 18TBX100, redrawn by eye (±1 dB). B&C don't state the drive level.
• Traced from B&C's S18H drawings (cm, converted to mm), 18 mm ply. The driver panel slopes 12.3°, and the throat under it opens from 70 mm at the front to 272 mm at the back wall.
• The same idea as the 18H, flipped: B&C put the mouth on the left in their drawing. Driver: B&C 18TBX100 (8 Ω). 598 mm wide (23.5 in).
• The back channel is open to the throat at the bottom: the sound leaves the driver, runs back under it and turns straight up. An earlier version of this drawing wrongly closed that corner. The white zone is a closed void between the paths.

• Chart: the TH18 page shows one builder's raw measurement ("El Phantasmo") in relative dB, redrawn by eye. It is not comparable in level to the other charts; the shape is the point. Peaks and dips above 150 Hz are why tapped horns need a steep low-pass.
• From the TH18 build sheet and its Hornresp record (diysubwoofers.org): areas S1–S5 are 268, 460, 1277, 2189 and 2718 cm², with segments 224, 2161, 320 and 205 mm. The build sheet labels the third segment 419; Hornresp says 320, and I drew Hornresp's.
• In a tapped horn the driver's front and back both feed this one path, at two different points. 18 mm birch, 571 mm wide (22.5 in). Drawn with a Precision Devices PD1850; built and reported with 18 Sound 18WL2400 and B&C 18SW115.
• Side section traced from the TH18 build sheet. The outer box, the 24.2° driver panel and the 508 mm mouth are dimensioned on the sheet; the inner fold panels are placed by eye from the drawing (±20 mm), so treat them as a guide, not cut dimensions. The front view shows where the mouth is: its 508 × 535 mm opening equals the horn's last area, 2718 cm².

• Chart: Rog Mogale's published 1 W/1 m plot with a PD1850, redrawn by eye (±1–2 dB).
• Traced from Rog Mogale's 2002 plan sheet. Rog co-founded Void, and this is the forum's usual stand-in for a Funktion-One F218. The driver fires up into a small chamber under the top, and the horn folds down and forward to a mouth in the lower front. The sheet doesn't spell out the air path, so the arrow is my reading. Panel corners are approximate to ±15 mm.
• Published: 50–200 Hz ±2 dB, 105 dB 1 W/1 m, 134 dB average / 137 dB peak, tuned 54 Hz, crossover 130–200 Hz. 600 mm wide (636 with sides). "Over 2 bins per side: House, Techno."
• Driver rule: PD1850, or any driver with EBP over 125 and BL over 28. Lower-EBP drivers go in his 186 Horn instead.
• License, verbatim from the sheet: "Feel free to build up to 4 cabs. If you want to make your living from this design then contact me for a licensing deal." Four is exactly our count.

• Chart: Rog Mogale's published 1 W/1 m plot with a P.Audio SD18, redrawn by eye (±1–2 dB).
• Traced from Rog Mogale's 2004 X1 sheets: 606 W × 770 H × 800 D mm, 18 mm 13-ply birch. The driver sits on a 540 mm baffle sloped 22.5°, loaded from a removable back panel. The open front is the 494 mm band between two 120 mm sections, with a centre brace.
• This is the least certain drawing on the page. The sheets don't say how the top and bottom 120 mm sections connect to the air path, so I've drawn them as solid. I'd read the X1 thread on his forum before trusting the arrow.
• Published with one P.Audio SD18: f3 31–119 Hz, 102.5 dB 1 W/1 m, 131.5 dB continuous / 135 dB peak, 800 W. It's deeper than the 1850 and gentler on output. Same 4-cab license.

• Chart: Tom Danley's 2002 predicted response at the mouth for one "short box" with the first-generation driver, redrawn by eye from the archived plot (±2 dB).
• Drawn from Tom Danley's dated 2002 sketch (panel names, open front) and the original 18 mm CAD file that the Live Audio Board published, recovered from the Web Archive. The panel positions come from the CAD; the air path follows Danley's labels. Dashed lines are open braces, not walls.
• How it works: two 12" drivers sit "push-push", each in its own small sealed back chamber, firing into one shared front chamber at the horn throat. The horn then folds around that module in a snail shell. Box: 45 × 45 × 22.5 in (1143 × 1143 × 572 mm), built for about 30–120 Hz and meant to be used in multiples.
• Known problems from the build threads: the back chambers came out about 14% small (21.5 L each), the drivers overheated in that small space, and builds showed a dip around 40 Hz. The original Eminence LAB12 driver is discontinued. Danley's DBH218 (2×18, 277 lb) keeps this same 45 × 45 × 22.5 footprint.
• It's a snail shell: the throat leaves the module upward between the module and the flare panel, turns over the top, runs down the back and along the bottom, and widens out the whole front face. The drivers fire across the width (into and out of the page), each from its own small back chamber against a side panel.

• Our design, from scratch. The idea comes from BASSBOSS's VS21: the back of the driver feeds a big ported chamber for depth, and the front feeds a short vent for punch. The numbers are ours. B&C 21SW152 driver; box 610 W × 967 H × 915 D mm (24 × 38.1 × 36 in), about 450 L inside.
• Simulated: one box at 2.83 V peaks around 102 dB. Max output is 133 dB at 40 Hz, 132 at 50, 133 at 63, 135 at 80 and 132 at 100. Below 35 Hz the cone runs out of travel. The rear port stays at or under 25 m/s air speed at full power.
• Bigger driver? In a bandpass the 21" earns its keep: about 2.5 dB more than the best 18" bandpass, but in a box about 35% larger. That's the right place to go bigger.
• Weight estimate: about 190 lb with the driver in 18 mm birch, so it needs handles and a wheelboard. Using poplar ply or an 18" driver brings it down.
• v0: panel positions follow the simulated volumes and lengths. Next steps are cross-checking in Hornresp, then building one prototype and measuring it against BYTE's subs.

• Tightened from v0. The geometry now drives the simulation: the three legs measure 217 + 672 + 643 mm = 1.53 m along the centreline, and those exact lengths and cross-sections are what's simulated. Throat 574 × 85 mm (a 2.5 : 1 squeeze on the cone), mouth 574 × 586 mm.
• Box 610 W × 924 H × 819 D mm (24 × 36.4 × 32.2 in), about 460 L outside, which is 7% smaller than v0. The magnet clears the top-leg floor by 22 mm against B&C's 242 mm driver depth. v0 would have hit it; my first clearance check looked at the wrong edge of the magnet.
• Simulated: about 102 dB at 2.83 V from 50–100 Hz. Max output is 134 dB at 50 and 63 Hz, 133 at 80 and 135 at 100, falling to 125 at 40 Hz where cone travel runs out. It needs a low-pass around 100–110 Hz.
• The rear chamber needs 76 L and the space holds about 101 L, so a vertical divider walls off the rest as a sealed void (good place for bracing).

• The extra fold. The horn wraps all the way around the driver block and finishes like the LAB horn: the last leg turns down into an open front zone, so the mouth is the whole front face. Legs: 406 mm along the floor, 535 up the back, 558 over the top, 214 down and out, so about 1.71 m. Throat 574 × 85 mm.
• Box 610 W × 865 H × 1008 D mm (24 × 34.1 × 39.7 in), about 530 L outside.
• Simulated: 136 dB at 50 Hz, 134 at 63 and 80, 137 at 100, and 127 at 40 Hz, which is 2 dB more than v1 down low. That's about 1–2 dB better than v1 for a box about 15% bigger. I tried two other four-leg layouts first: both lost mouth area to the driver and came out worse (about 128 dB).
• Magnet clearance to the block top is about 56 mm. The front zone is open, so the grille sits on the front edge.

• Every box seen from the front, next to a 5 ft 10 in person, with the mouth or port openings shown as air. Widths × heights × depths are in millimetres.
• The TH18 and LAB faces come from their plans: the TH18 mouth is the bottom half of the front, with the driver visible inside; the LAB has no front panel at all. The S18H and 1850 mouths are read from their drawings; the 1850's is approximate.
• FH18-W's dashed line marks the driver block, set back inside the open front.

• Our pick for the kick. Two B&C 12PS100 punch drivers stacked on one baffle, firing into a 50 mm front chamber and a 577 mm exponential horn. Throat 574 × 84 mm, mouth 574 × 558 mm. Sealed 43 L rear chamber.
• Box 610 W × 734 H × 819 D mm (24 × 28.9 × 32.2 in): the same width and depth as the FH18, so it sits flush on top. About 150 lb in 15 mm birch (estimate).
• Simulated: about 107 dB at 2.83 V across 100–250 Hz. Max output is 132 dB at 80 Hz, 137 at 100 and 125, 135 at 160–250. The band to use is about 85–220 Hz.
• Why two 12s: at the same box height they beat one 15" by about 4 dB across 85–200 Hz with a smoother curve, and two light cones are what gives the "snap". The cost is about 8 kg more driver weight.

• One B&C 15PS100 in the same kind of short horn: 498 mm long, throat 574 × 68, mouth 574 × 255. Box 610 W × 469 H × 817 D mm (24 × 18.5 × 32.2 in), about 115 lb (estimate).
• Simulated: about 102 dB at 2.83 V. Max 129 dB at 80 Hz, 131–132 at 100–160, 128 at 200. Given a full-height box (536 mm) it reaches about 130 dB with a lumpier curve.
• The 15PS100 is built for vented boxes (Qts 0.43), so it isn't ideal here; a horn-oriented 15" would close part of the gap. This is the light, cheap option, not the pick.

• Traced from Rog Mogale's 2001 plan sheet (all mm). Box 578 × 578 side, 472 wide (18.6 in): narrower than our 610, so it won't stack flush without a frame.
• My reading of the sheet: the front is the right-hand face, open below a 120 mm top band. One 15" sits on a 420 mm sloped baffle and fires forward through a short wedge-shaped horn. A curved panel (395 mm radius) closes the sealed rear chamber behind it. A removable bottom panel gives driver access.
• "Designed for Hard Dance Music. Use in stacks of at least 6 bins." "Offers bass from 52–180 Hz." Drivers, in his order: Precision Devices PD154, JBL E140, Eminence Kappa 15LF, Fane Colossus 15 Bass.
• License, verbatim: "Feel free to build up to 4 bins. If you want to make your living from these plese email me to arrange a licensing deal."
• Chart: his published 1 W/1 m plot (quarter space, so about 6 dB above our half-space numbers), redrawn by eye (±1–2 dB).

• The closest commercial match to K212: 590 mm wide, two 15s in "fully horn-loaded", vertically stacked horns, an operating band of 55/85–250 Hz, 102 dB at 2 V, 63 kg, 15 mm birch. Usually stacked on Funktion-One subs.
• Envelope only: Funktion-One publish the outside (590 W × 845 H × 603 D mm) but not the inside. The two mouths shown are approximate.
• Chart: the published band, not a measured curve.

• Void's "compact folded horn-loaded upper bass" box: 4 × 15", 60–190 Hz ±3 dB, 109.5 dB 1 W/1 m, 142 dB continuous, 4,000 W, 150 kg. 738 W × 748 H × 1218 D mm, birch with aluminium bracing.
• Void's own line on it: "The urge to dance predominantly comes from the upper bass frequencies, where the kick… [is] found." It is the benchmark K212 is chasing, at about a quarter of the weight.
• Envelope only; the interior isn't published.

• The archetypal kick bin: one 15" in a front-loaded horn with a vented rear chamber. "Usable response to 45 Hz with maximum loading above 200 Hz." 914 H × 762 W × 606 D mm, 41 kg without driver. Drivers: JBL 2220, 2205, E140 or E145.
• JBL's own note on the type: a front-loaded horn adds "6 dB to the on-axis sensitivity" in its range, and the vented rear chamber holds the response up "an octave or more below cutoff".
• Envelope only; clone plans circulate on forums but JBL's sheet has no section. The chart only marks JBL's stated range; the level is a placeholder.

• A DIY "port-assisted front-loaded horn" kick: 62 × 62 × 62 cm from one 4 × 8 sheet, 80–200 Hz with most output around 120 Hz, designed for the 18 Sound 15ND930. There are 12" (45 × 45 × 50 cm, 110–250 Hz) and 18" versions. The designer reports more than 10 prototypes.
• The designer's rule of thumb: one kick per two subs (1.5 : 1 for live music).
• Envelope only: the published cut sheets are thumbnail-size. No license text on the page, so ask the designer before building to rent.

• Widths × heights × depths in millimetres. Mouths are shown as air. For the reference boxes (F215, Hyperfold, JBL, Cubo) the mouth shapes are approximate, because only the outsides are published.
• K212 and K115 are 610 mm wide to stack flush on the FH18 and BP21.

Other references checked: BASSMAXX X2C "Deuce" (a 2×12 horn sub, 44–140 Hz, brochure) and Eighteen Sound's 2×15 "front load" note, which is a vented sub rather than a kick.

• Our design. A B&C 12MH32 (a 12" built for horn-loading: Qts 0.19, 54 g cone, 101 dB) on a short 258 mm horn from about 220 Hz to 1.2 kHz. Throat 574 × 120 mm, mouth 574 × 402 mm, sealed 29 L rear chamber. Above it, an Eighteen Sound ND3T 1.4" compression driver on a 60° × 40° waveguide (XR1464C) from 1.2 kHz up, the pairing Eighteen Sound use in their PS112.
• Box 610 W × 726 H × 485 D mm (24 × 28.6 × 19.1 in). It sits on K212 flush at the front. About 90 lb in 15 mm birch (estimate).
• Simulated mid: about 106 dB at 2.83 V across 200–800 Hz; max output 132 dB at 200 Hz, 132 at 300, 133 at 500, 132 at 800, falling to about 126 at 1.2 kHz where the compression driver takes over. HF from the datasheet: 112 dB at 1 W and 110 W rated, so about 132 dB before heat compression.
• That's about 8–9 dB more per box than the 12" two-way placeholder, which is the gap the configurator found. Crossovers: K212 → mid around 200–220 Hz, mid → HF at 1.2 kHz (the ND3T's recommended point, 12 dB/oct or steeper).
• Caveats. My horn model is one-dimensional, so above about 500 Hz, where a 574 mm-wide mouth starts to beam, treat the mid numbers as approximate; coverage needs measuring. The XR1464C needs a 240 × 240 mm cut-out and 180 mm of depth (Eighteen Sound's sheet). Next step: a horn with a horizontal flare too, so the mid pattern matches the 60° HF.

• John White's (JW Sound) DIY multiple-entry horn, the DIY descendant of Danley's Synergy idea (via Scott Hinson's MEH). JMOD 2.0 (plan rev 2.03, June 2026): two 12" (B&C 12NDL88; 18 Sound 12NTLW3500 also validated) plus a B&C DCX464 coaxial compression driver. 90° × 60°; pattern control down to 320 Hz across and 500 Hz vertically. 80 Hz–18.5 kHz, F3 55 Hz. 750 W × 470 H × 420 D mm, 34 kg.
• ⚠ It overlaps the kick. JMOD plays down to 55 Hz, across all of K212's 85–220 Hz band. Use it one of two ways: crossed to K212 at 200 Hz or below (JW Sound's own FAQ says to cross to kick bins no higher than 200 Hz), or instead of K212, straight over the subs, the way JW runs his own rig. Running both full-range doubles the kick band and smears it in time.
• At 750 mm it is 140 mm wider than our 610 mm boxes, so it overhangs a K212 stack.
• License, verbatim: "The JMOD is provided as a DIY-only design. Manufacturing, marketing, and distribution of this design for commercial purposes is prohibited under the … CC BY-NC-ND 4.0 license." Renting to friends is a grey area under "non-commercial"; ask John White. Danley's Synergy patent US 8,284,976 runs to Nov 2029 and may also apply.
• The side view is schematic: the plan has the real geometry but I haven't traced it. Max SPL is only a chart in the plan; the "129 dB" quoted online is unverified.

• Eighteen Sound's app note (and the CNC files we have): a 12NTLW3500 woofer (97 dB, 900 W) in a 50 L box tuned to 58 Hz, with the ND3T on an XR1464C (60 × 40) or XR1496C (90 × 60) horn. 378 W × 680 H × 419 D mm, a trapezoid with 10° sides, 15 mm birch.
• Preset: woofer low-passed at 1.4 kHz (LR24), compression driver high-passed at 1.8 kHz. Coverage about 90° at 1 kHz, narrowing to about 60° above 3 kHz.
• About 123 dB continuous per box by my estimate, which is why it was the configurator's placeholder and why it comes up short. It's the right box for DJ monitors and fills.
• Front view shows the horn, woofer and port slot at their approximate positions. Terms: "© Eighteen Sound Srl … all rights reserved", no build license stated.

• The template for MH12. Funktion-One's skeletal mid-high: a 10" on a horn from 220 Hz to 4 kHz plus a 1.4" from 4 kHz, in an open frame. 605 W × 618 H × 596 D mm, just 23 kg, 40° × 20°. Its width is within 5 mm of ours.
• Funktion-One crossover points: 85–220 Hz (bass), 220 Hz–4 kHz (mid), 4 kHz up (HF). They run the 10" much higher than we plan to.
• Envelope only, mouths approximate.

• Void's skeletal open-frame horns: 12" low-mid, 8" mid, 1.5" HF, each on its own isometric horn. 854 W × 672 H × 658 D mm, 35.4 kg, 60° × 50°, 134 dB continuous / 137 dB peak. DC-10 Ibiza runs eight of them.
• The look we could borrow for MH12: the horns as visible structure, with coloured bracing. Envelope only.

• Danley's narrow Synergy horn: 2 × 12", 4 × 4", one 1.4" in one horn. 567 W × 737 H × 577 D mm, 53.5 kg, 40° × 60°, 135 dB continuous / 141 peak, the loudest box per kilo in this list. Patented (US 8,284,976, to Nov 2029).
• It shows what a multiple-entry horn can do at our width. Envelope only.

• Celestion's free build-it-yourself kit: TF1225 woofer plus CDX1-1745 on an H1-9040 (90 × 40) horn, crossed at 1.9 kHz. 388 W × 654 H × 320 D mm. 123 dB continuous / 129 peak.
• The 15" version (CKT-TF1525e, 452 × 730 × 342) reaches 124 / 130. Both confirm the ceiling of 2-way kit tops at about 123–124 dB. Envelope only.

• Our idea. Two B&C 12MH32 mids, each lying flat at the back and firing toward the outer wall, so each one's sound turns and runs forward through a narrow fold (50 mm high, 215 mm long) and then a straight flare (242 mm) to its own slot. One slot sits above the HF waveguide and one below, so the box is symmetric top to bottom. HF: ND3T on a 60 × 40 waveguide in the middle.
• The fold is kept narrow (50 mm, well under half a wavelength at 1.2 kHz) so it doesn't act as a filter; all the widening happens in the straight run to the slot.
• Simulated: about 113 dB at 2.83 V across 200 Hz–1.2 kHz with both mids, and 137–140 dB max on axis. That's about 6 dB above MH12, because there are two mids.
• ⚠ The catch: the slot centres are 476 mm apart, so above and below the axis the two mids start to cancel. The first cancellation is at ±17° at 1.2 kHz, inside a 40° vertical pattern. Crossing to the HF at 800 Hz instead (which needs a 2" driver such as B&C's DE750) moves it to ±27°, outside the pattern.
• Patent check result: low risk. The mids load their own horns and their slots exit outside the HF waveguide, so the element Danley's Synergy claim needs, a lower driver "mutually coupled" into the HF horn's passageway, appears to be missing. Keep it that way: slots open at or in front of the HF mouth plane, never into its flare. The nearest active patent is Meyer Sound's US 10,015,583 (to 2036), which covers direct-radiating mids angled inward at the HF horn's sides, venting beside and above and below the mouth. FPS-S's horn-loaded folds don't do that. Not legal advice; worth a quick freedom-to-operate check before renting.
• Box about 610 W × 712 H × 650 D mm (28.0 × 25.6 in). Concept layout: lengths and areas match the simulation, but it isn't cut-ready.

• Our idea, in its purest form. One horn, 60 × 40, with the compression driver at its apex. Two 12MH32s face each other across the width (push-push, like the LAB horn) into one shared chamber. Their sound travels through a short fold and enters the horn through slots in its top and bottom walls, near the throat. Mid and high come out of one mouth, from one point: Danley's point-source result, fed through a fold instead of cones mounted on the horn's sides.
• Simulated: about 112 dB at 2.83 V, flat within about ±0.2 dB from 220 to 500 Hz, and 136–139 dB max. Entry 171 mm from the HF throat; one mouth of 574 × 362 mm.
• ⚠ Physics sets the crossover. For mid and high to act as one source, the mid has to enter within about a quarter wavelength of the HF throat. Two 12" cones squeezed into the horn that close to the throat compress the air hard: 4.3 : 1 at a 500 Hz crossover, 7 : 1 at 650 Hz, 10.8 : 1 at 800 Hz, 17 : 1 at 1 kHz. My model is linear, so it can't show the distortion that heavy compression causes. Hence the 500 Hz crossover, which needs a large-format compression driver rated that low. It's also why Danley uses several small mids instead of two big ones.
• ⚠ Patent check result: high risk in the US until about Nov 2029. Claim 1 of Danley's US 8,284,976 needs only an HF driver at the first end of a horn passageway, a lower driver "mutually coupled" to that same passageway, and the lower driver kept below its "first cancellation notch". There's nothing about wall ports, distances or mounting, so feeding the mids through a fold does not obviously avoid it. Dependent claim 4 even covers coupling "by an aperture in the sound barrier". The European and Canadian versions have reached their 20-year term; the US one runs to about 2029-11-06. In the US, personal or DIY use is not a defence, and renting to friends certainly isn't.
• Ways forward to discuss with a patent attorney: (1) invert it: make the mid fold the main horn, with the HF waveguide nested in its mouth (expired Harman US 7,236,606 is prior art for that); (2) let the mid exit at the HF mouth plane instead of through the walls, which moves it toward FPS-S; (3) ask Danley Sound Labs for permission; (4) wait for expiry in late 2029. Useful expired prior art: Renkus-Heinz CoEntrant US 5,526,456 (1993) and Danley's own Unity US 6,411,718 (expired).
• Box about 610 W × 496 H × 863 D mm (concept). The side view is schematic about the mids: they sit across the width, so here they appear face-on.

• David's design. One 60 × 40 horn (the cone) with the compression driver at its apex at the back. Two 12MH32 mids lie flat in the front corners, top and bottom. Each feeds a full fold: back along the outer wall, round the back corner, then diagonally inward to an opening in the cone's wall 172 mm in front of the HF throat. Both walls of each fold land on the cone at the edges of its opening, so everything behind the openings, around the HF, is a sealed space. Mid and high leave through one mouth, 574 × 362 mm.
• The math: each fold is about 870 mm of 26 mm-high channel across the full 574 mm width, widening slightly into the opening. Simulated: 110–113 dB at 2.83 V, 136–139 dB max, 2.1 dB ripple from 220 to 500 Hz, 8 m/s air speed in the folds at full output. The long fold costs almost nothing compared with a short duct (2.0 dB ripple).
• Crossover 500 Hz: the mids must enter within a quarter wavelength of the HF throat (172 mm at 500 Hz), which keeps the squeeze about 4 : 1. That needs a 2"-exit large-format compression driver rated to 500 Hz: the Radian 950PB (4" aluminium diaphragm, 4" edge-wound voice coil, 500 Hz–20 kHz, 113 dB at 1 W on a horn, 125 W AES, 4.7 kg, about $950; a beryllium 950NeoPB also exists). One driver covers everything above the mids, so about 134 dB before heat compression, matching the mids. Cross at 500 Hz with a steep slope (24 dB/oct) and a limiter, because 500 Hz is the bottom of its range. Its 4" diaphragm will roll off a little at the very top, so expect a gentle shelf EQ above about 10 kHz. Rejected: JBL D405 (a midrange compression driver that stops around 7 kHz) and B&C DE950 (2" exit but rated to cross no lower than 1.2 kHz).
• Each mid wants about 30 L sealed behind it; join the front-corner chamber to the void behind the fold to get it. Box about 610 W × 786 H × 676 D mm (24 × 31 × 27 in). Concept, not cut-ready. The mids lie flat (cones at the top and bottom walls) to keep the height near 31 in; forward-facing corners would make it about 1.1 m.
• ⚠ Patent: the mids enter the HF horn, so this reads on Danley's US 8,284,976 claim 1 (to about Nov 2029 in the US). FPS-2.2 below is the design-around.

• The design-around, same look as 2.1. The two folds now feed the mid horn from its throat at the back centre: each runs back along its outer wall, down the back wall, and the two meet at a 51 mm throat. The mid horn flares forward to the 574 × 362 mm mouth. The HF compression driver and its 60 × 40 waveguide sit in a streamlined pod nested in the mouth, so the mids flow around the pod on all four sides, and mid and high still leave the same mouth from the same centre.
• Simulated mids: about 110–112 dB at 2.83 V and 132–138 dB max across the whole 220 Hz–1.2 kHz band (2.1 only runs to 500 Hz). The long path rolls off about 6 dB by 1.2 kHz; a gentle shelf EQ or a crossover nearer 1 kHz fixes it. The pod takes about 28% of the mouth; the model includes that.
• The big win: the HF has its own waveguide, so the mids don't have to enter near its throat. That means no 500 Hz crossover and no large-format driver: the regular ND3T crosses at 1.2 kHz (about 132 dB from its datasheet).
• Vertical coverage: the mids leave 61 mm above and below the 240 mm pod, centres 301 mm apart, so their first cancellation is at ±28° at 1.2 kHz, outside the 40° pattern (FPS-S was ±17°). Across the width the mids wrap the pod on both sides too, which tightens it further.
• Patent: the mids never enter the HF waveguide, so the "mutually coupled to the horn passageway" element of Danley's claim appears to be missing. Expired prior art for an HF horn in the mouth of a mid horn: Harman US 7,236,606. Low risk, like FPS-S; still worth one attorney check before renting.
• The pod (simulated): its front face is flush with the outside of the box, in the mouth opening, with a wedge tail tapering to a point 450 mm back, like a W-bin splitter. The sound coming up from the throat meets a knife edge instead of a flat wall, and the open area around the pod only ever grows: zero pinch and 6.4 dB ripple, the same as no pod at all (a blunt flush pod was 15% pinch and 6.8 dB). The 450 mm length is set by the HF driver: the ND3T sits about 240 mm behind the waveguide mouth, where the wedge is still 112 mm tall. Tails from 250 to 550 mm all have zero pinch; shorter ones can't hold the driver, longer ones squeeze the gap toward the throat. Pushing the pod deeper with its blunt back was the worst option (8.1 dB at 100 mm, and it hit the walls).
• The HF is sealed at its source. A compression driver is a closed unit (its diaphragm's back is enclosed in its own rear cap), so the wedge only houses it; there is no rear chamber or vent. The wedge shell must be airtight: glued and sealed panels, the waveguide flange gasketed to the wedge front, and the HF cable through a sealed gland in a strut. Otherwise mid-horn pressure leaks into the cavity and colours the sound. It also matters for the patent: the HF throat connects only to the compression driver, and the mids never enter the HF passage. Line the cavity with damping, use stiff 15–18 mm panels (the wedge is pressure-loaded from outside), service the driver from the front by unbolting the waveguide, and leave a few millimetres of air around the driver's rear cap for heat.
• Wiring the floating HF: the pod hangs on a thin vertical fin (about 18 mm, on the centre line) above and below it, near the front where the vertical gaps are widest. The fins line up with the airflow and block only about 3% of the horn's width, and most of the mid sound passes the pod's sides anyway. The HF cable runs through a channel routed inside the upper fin, through the horn wall into the sealed void, then to the jack panel. Use a sealed gland at each end so the pod and the void stay airtight. Round the fins' leading edges (toward the throat) and taper the trailing edges so they don't whistle. Leave slack and a disconnect in the pod so the driver can come out through the front with the waveguide.
• Most of the mid actually passes the pod's sides: it's 240 mm wide in a 574 mm horn, so there's about 167 mm each side at the mouth. Box about 610 W × 786 H × 676 D mm. The mouth is a 574 × 362 mm opening cut through the front panel; the horn walls run through the panel to the front face, the wedge's front (the HF waveguide mouth) sits flush in that opening, and the mids come out of the same opening around it. Nothing sits in front of the wedge except a grille. Concept, not cut-ready.

• Our folded point-source concepts are drawn to scale on Plates T8–T11 (FPS-S about 610 × 712 × 650, FPS-1 about 610 × 496 × 863, FPS-2.1 and FPS-2.2 about 610 × 786 × 676 mm).
• Mouths shown as air; for reference boxes they are approximate. MH12 is 610 wide to sit on K212; JMOD (750) and Air Motion (854) overhang a 610 stack.

Also checked: Rog Mogale's mid-top plans on speakerplans.com (X10, X12, X15 and others; image-only, no license text found), Bill Fitzmaurice DR300 and OmniTop, B&C BOX12/BOX15 (PDFs blocked this time).