Mono 5-way plus piezo fan. DCX2496 -> off-the-shelf XLR -> 4 × QSC GX5, knobs FULL UP. Build sheet V3, rev n (2026-10-21), after the driver swap: scoops are now JBL 2226H, kicks EV EVM-15L.
The horn limiter value assumes the −20 dB pad is in. See page 2-5.
Gains set by ear 2026-10-22, trimmed 2026-10-23 with REW (horn up 2 dB to meet the house curve at 1-4 kHz). The paper model (scoops −7.5, kicks −8, horn +11.5) was far too bright. Keep the subs about 4.9 dB under the scoops: if the scoop gain moves, move the sub gain with it.
Every connection from the DCX to the amps is an off-the-shelf XLR cable. Each DCX output is a female (F) jack; Y-cables split one female into two males (M) for the pairs.
Hookup wire runs from each amp to the 8-hole Speakon panel on the back of the rack; Speakon cables go from there to the boxes.
Fig. 2-2. Amplifier outputs to the speakers.
AT THE AMP
Best: a Speakon plug (NL4FX) on the amp end (the GX5 jack uses pins 1+ and 1-). It locks and cannot shake loose.
Or the binding posts: crimp a fork terminal or ferrule on, or strip 1/2 in, twist tight and do NOT tin it. No stray strands. Re-tighten after the first loud sessions.
COLOR CODE
Each band has one color, marked on the amp output, the cable at both ends, and the input of the box. Match color to color.
The horn channel has a -20 dB inline XLR pad (PE 240-416) between the DCX and the amp, plugged straight into GX5 #4 input B. It does two jobs: it lets the horn's limiter be set at all, and it keeps the horn from hissing.
WHY: THE LIMITER
The 2440 is very sensitive (111 dB from 1 W) and rated for only 60 W into 16 ohm. That is about 31 V at the speaker. With the amp knob fully up (x52.5), the DCX would only have to send 0.59 V, -2.4 dBu, to reach it: a limiter setting of -24.4. The DCX's limiter only goes down to -24, so without the pad the horn cannot be limited at all.
With the pad in, the DCX sends 20 dB more (+17.6 dBu) for the same 31 V, and the limiter sits at -4.4, comfortably inside its range.
WHY: THE HISS
Every DCX output carries a little noise, about -90 dBu, all the time. The amp turns that into hiss. On the bass channels it is inaudible, but the horn turns very little voltage into a lot of sound:
DCX noise, no padabout 41 dB at 1 m
DCX noise, with padabout 21 dB at 1 m
GX5's own noiseabout 35 dB at 1 m
Quiet roomabout 30-35 dB
Without the pad the DCX's hiss is clearly audible from the horn in a quiet room. With it, the DCX noise drops 20 dB, below the amp's own noise. The pad turns the music down too, but the DCX now runs 20 dB hotter to make up for it, so the music stays as loud and the hiss is gone. What remains is the amp's own faint hiss, about as quiet as the room. (Estimates from the DCX and GX5 specs and the 2440's sensitivity.)
CHECK THE PAD
Ohmmeter: pin 2 to 2 and 3 to 3 read the same, not zero; 1 to 1 = 0 ohm. Pin 3 at 0 ohm means a one-leg pad: return it.
Measure it (7-3). Label it. It never moves. Never run the horn without it.
All GX5 knobs fully up. This is deliberate. The limiter settings are calculated for the amps' full gain, so the limiters cap what reaches every driver. With the knobs at the top, nobody can turn an amp UP past that point: an eager knob-turner can only make a band quieter, never push it past its limiter. If the knobs sat lower, turning one up would add gain after the limiter and bypass the protection. All level changes happen at the mixer or in the DCX (PAGE LOCK on).
Limiters are the ONLY protection. Set them all before playing anything loud.
Limits by math: DCX 0 dB = +22 dBu; OUT GAIN is before the limiter, so trims don't move thresholds.
Subs + scoops limit together: subs at 150 W (half rating: old drivers, one cone repaired). Scoops (2226H) at 465 W, the GX5's limit; the horn model puts them at 76-91 % of Xmax there. Subs gain sits 4.9 dB under the scoops, the gap between the two limiters (−23.4 vs −18.5).
Kicks (EVM-15L) at 200 W, the driver's rating, which is also close to its Xmax above the 80 Hz high-pass.
Check: pink noise on In A, both outputs' limiter LEDs light at the same input level. Trim bass vs mains at the mixer, not here.
SUBS: 300 W OPTION
The label's 600 W is peak; about 300 W continuous is the ceiling. Stay at 150 W until a loud test shows both:
Out 1's LIMIT LED lights, with about 35 V AC at a sub's terminals (150 W into 8 ohm).
Both subs stay clean at the limit, the repaired cone included: no rub, buzz or flapping.
Limit300 W into 8 ohm, about 49 V
Limiter+1.6 dBu = DCX −20.4
Limit together: sub gain 1.9 under the scoops (+6.1 with scoops at +8). Heavier bottom: keep it 4.9 under; the subs then carry on 3 dB past the scoops.
Set so a loud record with the mixer's masters at nominal just reaches the bass limiters. Below nominal: drops and swells. Above it: the limiters catch.
SETTING
IN A gain−15 dB
IN B gain−15 dB, the same as IN A
From the mixer¼" PGM OUT 1 and 3, not RCA
Factory default0 dB
THE MATH
IN gain = limiter − OUT gain + 22 − mixer peak (dBu).
Out
Band
Limiter
OUT gain
Limits at, IN 0 dB
1
Subs
−23.4
+3
−4.4 dBu
2
Scoops
−18.5
+8
−4.5 dBu
3
Kicks
−22.1
−2
+1.9 dBu
4
Mid
−18.6
0
+3.4 dBu
6
Horn
−4.4
+8
+9.6 dBu
A loud record at 0 VU peaks at about +10 dBu, so IN A wants −14.5; −15 (the DCX's bottom) limits the bass at +10.5. With IN B at −15 the kicks limit at +16.9, under the mixer's +20.2 maximum: master 3 has about 6 dB more travel.
Keep IN A = IN B. The OUT gains hold the tuning, so equal master faders give the reference setting on page 2-13.
Recompute when an OUT gain or limiter changes: 1 dB more on the scoops = 1 dB less on IN A.
Check: masters at nominal, the LIMIT LEDs on Outs 1 and 2 just flicker on loud peaks (a band carries less than full range, so a little later than the math).
Horn on Out 6 -> limiters -> cables -> each channel alone, quiet (right box? polarity?) -> delays -> trims -> store -> PAGE LOCK (write the password down).
SUB HIGH-PASS
40 Hz LR48, at the measured port tuning (about 39-40 Hz). Raise it only if the cones flap or bottom out (Section 3).
Fig. 2-3. Stack, front view (not to scale).
Bottom row: scoop, two stock Ruby subs stacked, scoop; 24" x 9" platform boxes on the scoops, lettered MAS and HUP, make a level top with the subs (about 50"). Kicks lie flat across it, side by side, touching in the middle. The mid horn sits centered on their seam with the 2350 on top, fronts flush, and the piezo plate on the 2350. Strap the top. Corner blocks locate the bottom row; every box stays separate.
Speaker: 8 Speakon runs, about 12 ft of 13 AWG each, + 1 horn -> piezo jumper.
Every speaker has a Speakon inlet.
Box polarity, 9 V battery on each box's Speakon, + to 1+: every cone moves FORWARD. JBL drivers (2226H, 2123H, 2440) are opposite to EV: + on the BLACK terminal is forward. Check every scoop and kick after the driver swap.
DCX Auto-Align makes every band arrive at the listener together. Measurement mic into DCX INPUT C. Re-run whenever the stack changes.
Stack in its final position. Alignment is geometry: re-run it whenever boxes move, at a new venue, and whenever a box or driver is changed.
Crossovers and limiters set first (pages 2-1 and 2-2); crossovers change the timing. Horn pad in place.
Polarity NORMAL on every output. Delays at 0, DLY ON per output, LINK OFF (it is global).
One box per pair while measuring: turn the B knob fully down on GX5 #1, #2 and #3 (Sub 2, Scoop 2, Kick 2) so the mic hears one clean arrival per band, not two. The pair still shares one output, so the delay the DCX writes applies to both boxes.
Quiet room, doors shut, no one between mic and stack.
SET UP THE MIC
Measurement mic -> DCX INPUT C (XLR). Turning AUTO ALIGN on switches C to mic level and turns phantom on.
Place it centered on the stack, about 3 m out, at ear height (5-5½ ft), pointing at the stack. Not closer than 2 m: too close and the nearest box wins.
SETUP -> DLY-CORR./AUTO-ALIGN: DELAY CORRECTION ON, set the room temperature.
Sanity check the numbers. Expect a few ms at most (SHORT max 11.64 ms, about 4 m). A value near the limit means a bad measurement: move the mic and re-run. Measured 2026-10-23: the scoops arrive last and get 0 (their folded horn is the longest path); kicks get the most, 3.88 ms. An earlier run (2026-10-22) had the bass the other way round: auto-align is unreliable below 100 Hz, so check the bass with REW (page 2-14).
Re-run once with the mic 1 m toward the boxes still playing (B knobs still down). Values within about 0.1 ms: keep the first set. Big swings at 45-80 Hz are normal (long wavelengths); between mid and horn they are not.
B knobs on GX5 #1-#3 back to FULL. The limiter settings assume every knob is fully up.
Pink noise from the mixer at a comfortable level, all bands on.
Polarity test at each crossover, after delay, never before:
Mid <-> horn (700 Hz): flip Out 4. The 700 Hz region should go hollow. Flip back: full = correct.
Kicks <-> mid (300 Hz): flip Out 3, listen for the dip, flip back.
Scoops <-> kicks (80 Hz): flip Out 2, same test.
Subs <-> scoops (same band): flip Out 1. The bass should thin out badly; back to normal = correct.
If a seam sounds fuller inverted, leave it inverted and note it, but suspect the timing first.
Store the preset as MASHUP and the next number (MASHUP 1, MASHUP 2, ...; the name field is too short for a date), then PAGE LOCK. Log the number, date and what changed.
Fig. 2-4. Measured with REW 2026-10-23, 1/6-octave smoothed. Thin lines: each band alone, mic 10 ft out on the stack centre line. Thick line: the whole system, averaged over three mic spots (10 ft centre, 10 ft left, 20 ft back). Dotted lines: crossover points.
The reference setting on page 2-13, studio, sweep at −12 dBFS. Every crossover adds: at each seam the whole system is louder than either band alone.
Bass: about +10 dB at 60 Hz over 500 Hz, a heavier bottom than the JBL/Harman house curve (about +5), on purpose. Below 45 Hz the sub high-pass takes over.
1-4 kHz: on the house curve. Above 6 kHz the 2440 rolls off; that softer top suits dub.
The room: at 10 ft centre there is a 14 dB dip at 150 Hz (see the kick line). It moves with the mic, so the average shows only a sag. Room, not settings: do not EQ it.
