Step 2 — Port tube

Question: where should the damped box be tuned for the Aiyima, and how long must the tube be? Result: a 175 mm port moves the tuning from ≈ 72 Hz to ≈ 48 Hz; the +8 dB bass hump turns into a response flat within ±1 dB from 55 to 150 Hz.

Target

• After damping the bare 44 mm port tuned to ≈ 72 Hz, with a hump at ~79 Hz and a steep fall below 58 Hz.
• Target ≈ 50 Hz. The original design tunes to 43 Hz for the lean FE103NV2 plus a nearby wall; the Aiyima has more bass of its own (Aiyima vs. FE103NV2).
• Constraint: 188 mm inner depth. 43 Hz would need ≈ 216 mm of port; too close to the back panel the air at the tube's mouth gets choked.

Prediction

The tube length is scaled from the measured tuning (Helmholtz: Fb ∝ 1/√L_eff, with end correction), so no box model is needed. From Fb = 72 Hz at 44 mm:

Target Fb	Total port
55 Hz	112 mm	fuller, less deep
50 Hz	146 mm	the target
≈ 46.5 Hz	175 mm	fitted — cut long, to trim towards 146 mm
43 Hz (plan)	216 mm	doesn't fit the 188 mm depth

Build

HT DN75 drain pipe, mounted from the inside against the front panel, joint sealed. Total port = tube + front panel + doubler = 175 mm, measured straight through. ≈ 4.7 cm clear to the back-panel wool. Details: Build → Port tube.

Verification

Four independent checks, then the ears:

1. Impedance (L-z-02; L-z-03 with all seams and the tube base taped, to rule out leaks).
2. Port near field — where the port's output peaks is Fb.
3. Bass sum (cone + port, complex, area-scaled) — its shape decides whether the tuning is right.
4. Far field at 50 cm — gated (direct sound) and 300 ms (in the room).
5. Listening A/B — left (tube) against right (same damping, no tube) with equal-level test signals.

Reference control — two gain knobs had moved between sessions. The loopback gain was reset with a tone sent only to the loopback (back to −22.3 dBFS, as in every earlier session); the mic gain was matched so the cone's output at 100–200 Hz equals the pre-tube session within 0.5 dB — a range no port can change. Far-field fingerprint afterwards: mic −10.9 dBFS (was −11.6), loopback −22.2 dBFS, arrival 1.63 ms.

Results

The speaker now, 30 Hz–20 kHz, room-free:

Below 400 Hz the near-field cone + port sum, above it the 3.4 ms gated far field, level-matched at the splice. From ~45 Hz to 4 kHz within about ±4 dB; the bass hump is gone. Two things remain: the treble peaks at ~6 and ~10.5 kHz (for a correction EQ), and the near-field part is half-space — free standing, the bass sits a few dB lower relative to the mids (baffle step), which a nearby wall gives back.

Bass, cone + port, levelled to 150–250 Hz: with the 175 mm port within ±1 dB from 55 to 150 Hz, −3 dB at 47 Hz, −7 dB at 40 Hz — "flat down to a point, then a clean roll-off". The length is right; it stays.

Port output: the peak — the tuning — moved from 71 to 47.6 Hz; the port delivers ~6 dB more at 40 Hz and leaves 70–100 Hz to the cone.

In the room (50 cm, 300 ms): the 70 Hz peak drops from +10 to +3 dB, 40–50 Hz rise by 6–12 dB. The room still shapes the bass (rear wall 50 cm behind the speaker) — that's for placement, not the tube.

	Damped, bare hole	175 mm port
Port tuning Fb	≈ 72 Hz	≈ 48 Hz
Bass hump	+5.9 dB at 79 Hz	none (max +1.0 dB)
−3 dB point	58 Hz	47 Hz
Bass sum at 40 Hz	−18.6 dB	−7.1 dB
Direct sound above 300 Hz	unchanged within 1 dB

Impedance: the upper peak grows and moves from 93.8 to 91.1 Hz, as expected for a lower tuning. The lower peak and the dip that mark Fb shrink to a plateau at 45–50 Hz: the port resonance is heavily damped (the fill, and wool close to the tube's mouth). Taping every seam changed little, so it isn't a leak. Fb is therefore read from the port near field (47.6 Hz) and confirmed by the plateau (Impedance jig).

Listening

Equal-level test signals sent to one speaker at a time — a bass staircase of single notes from 40 to 120 Hz and a slow sweep 30–200 Hz — left speaker (175 mm port) against right (same damping, no tube yet).

"Not a huge bass addition, but the speaker sounds more balanced in some way."

That's the measurement in words: the tube doesn't add level, it removes the hump at 70–80 Hz that read as "lots of bass" and spreads the output evenly down to ~47 Hz.

Conclusion

• The first cut landed on target. Partly luck — but the prediction came from the measured tuning, and three independent checks (port near field, bass sum, impedance plateau) agree on ≈ 48 Hz.
• Measurement makes it verifiable, the ears make it final. Tuning further would trade evenness against bass quantity — a matter of taste.
• Balance, not quantity. For more bass level: placement closer to a wall (as the original design intends), a gentle EQ shelf, or loudness compensation.

Open

• Port noise at high level (only 4.7 cm to the back-panel wool): a loud listening test decides; if it chuffs, trim 10 mm (Fb +1–2 Hz).
• Protective high-pass in the DSP lowered from 56 Hz to 36 Hz (0.75 × Fb).