DIY Hi-Fi Loudspeakers
A custom-built 2-way active loudspeaker pair designed from scratch, integrating acoustic simulation, diffraction modeling, modal FEA, hand-tool construction, power handling, driver selection, active crossover design, frequency-response testing, and system tuning. The project focuses on high-fidelity sound reproduction on a $500 budget.
VituixCAD
DSP/SigmaStudio
Power Handling
Acoustic Design
Power Tool Fabrication
SolidWorks
Simulation (FEA)
WinISD
DIY Electronics
In Progress
Design Goals & Constraints
Budget
$500
Use Case
Living room, flanking TV
Sensitivity
≥85dB at 2.5m
The speakers needed to deliver wide, natural dispersion for TV and music listening at the couch position while hitting these output and extension targets within the power budget. This was achieved through heavy electroacoustic simulation, ensuring a $500 budget would not compromise performance. How each constraint was addressed strategically is covered in the sections below.
Construction
Hand tools only, no CNC
listening Distance
~2.5m (couch)
Volume
16.0L (vented)
Topology
2-way active (biamped)
F3 Target
Below 40Hz
Crossover
~1.8–2.5kHz
Power
4x100W amp, DSP-controlledÂ
Air Velocity
<20 m/s at 20W
Ripple Target
±2dB above tuning
Electroacoustic Design
Driver Selection
The woofer is a Dayton Audio DC160-8, a 6.5" 8Ω driver chosen for its strong measured performance and value in its price range. The tweeter is a Dayton Audio DC28F-8, a 1-1/8" silk dome with 8Ω impedance and 89dB sensitivity. Together they fit comfortably within the $500 system budget while supporting the F3 and sensitivity targets defined in the design goals. Their overlapping frequency ranges and dispersion characteristics provide a solid foundation for the crossover.
Amplification & Crossover
The system is powered by a Dayton Audio KABD-4100, a single board that provides four independent 100W channels (at 6 ohms), along with a built-in Bluetooth receiver. An onboard ADAU1701 DSP, programmed in SigmaStudio, enables a fully active crossover where each driver has its own dedicated amplifier channel and independently tunable filter. This architecture eliminates passive crossover components and gives direct control over crossover point, EQ, and time-alignment correction.
Amp, DSP, bluetooth receiver combination board
Enclosure Volume & Tuning
WinISD modeling landed on 16.0L net internal vented enclosure volume and 42 Hz tuning frequnecy as the alignment that best met the design goals. This volume supports an F3 below 40Hz while keeping response ripple within approximately ±1.5dB above the tuning frequency, balancing bass extension and a smooth response.
Power Handling & Listening Level Analysis
- Calculated the musical crest factor for typical listening material, which puts peak transient power demand around 49W against a realistic continuous/average listening level of about 3W, so 3W was treated as the real-world continuous operating power, not the amp’s 100W rating.
- At 3W, modeled single-driver output is roughly 82dB at 2.5m; with stereo summation this reaches about 85dB at 2.5m, meeting the output goal.
- Checked both the 3W continuous case and the 49W crest-factor peak case against port air velocity and woofer cone excursion limits. 20W was the more realistic max power to check for when considering musical crest factors due to power being split over the frequency range.
- Both port velocity and cone excursion came out over target, but only at the rare ~50W peak case, not at the 3W continuous level.
- Decided that if something has to give at that peak, excursion running slightly over is the better failure mode than the amp clipping, a standard, accepted tradeoff since hitting full peak power is a brief, infrequent event in real listening.
- Addressed the port-velocity side of that tradeoff by adding a flare to the port, which bought back enough headroom to keep velocity reasonable even under peak conditions.
SPL level at 3W continuous signal. Graph shows around 82 dB before low frequency rolloff
Cone excursion at 20W signal. Clears the excursion limit with acceptable peak at ~55 Hz. High pass filter will be added through DSP to address below ~38 Hz (f3)
Port air velocity graph at 20W signal. 21 m/s peak is acceptableÂ
Port Design Tradeoffs
- Started at a 7.73cm diameter (3" PVC) port to minimize air velocity, but it did not physically fit inside the enclosure even after modeling a bent/elbow path in CAD.
- Reduced the port diameter to 4cm to solve the fit problem (smaller diameter = shorter length), but this pushed air velocity too high at peak power.
- Landed on a 2" diameter port that fit the enclosure and kept velocity around the 20 m/s target at 20W, but still went slightly above target air velocity at 50W peaks.
- Added a flared port end to the 2" design, which raised the acceptable velocity ceiling to roughly 30 m/s (versus about 17–20 m/s unflared), specifically to cover 50W peaks.
- Port diameter and length cannot be optimized independently, they jointly determine tuning frequency, low-frequency extension, and air velocity, so each diameter change required re-solving for port length to hold the same tuning point.
Acoustic Simulation: Diffraction
Baffle diffraction effects were simulated in VituixCAD to decide tweeter placement before finalizing the baffle layout.
Centered Tweeter Diffraction (blue line)
The Fix: Offset Tweeter (red line)
With a tweeter centered on the baffle, both edges are exactly equidistant, producing sharp and pronounced notches and dips in frequency response. The centered configuration’s on-axis response dropped about 6 relative dB at roughly 2.3kHz. This notch is followed by a pronounced secondary resonant bump in the 3.5–4kHz region.
Offsetting the tweeter breaks that symmetry, creating individually shallower frequencies instead of stacking into one deep notch. In the offset simulation, the on-axis response only dipped around 2 realtive dB in the same region, with smaller, smoother ripples afterward. The sharp secondary resonance seen in the centered case was effectively removed.
Off-Axis Symmetry
Because a centered tweeter is symmetric, its ±30° (red) and ±60° (green) off-axis curves overlap almost exactly, there is no assymetry between the positive and negative 30 degree line, for example. There is no directional cost, but also no on-axis benefit. Offsetting introduces roughly 1–2dB of asymmetry between the near and far ±30° and ±60° off-axis responses in the 2–5kHz band.Â
On-axis tweeter diffraction simulation with ±30° (red) and ±60° (green) off-axis curves
Off-axis tweeter diffraction simulation with +30° (green) and -30° (red) curves. Assymetry between plus and minus curves is kept between 1-2 dB
Off-axis tweeter diffraction simulation with +60° (green) and -60° (red) curves. Assymetry between plus and minus curves is kept between 1-2 dB
Decision
The offset configuration was selected because it directly improves the on-axis response in exchange for 1-2 dB of assymetry in the off-axis responses, rather than ~6dB dips in the centered tweeter, on-axis response. The tweeter was offset inward, toward the room’s center.
Modal FEA Analysis
SolidWorks Simulation modal (frequency) studies were run on the enclosure to screen for panel resonances and validate the internal bracing design. Three bracing configurations: no bracing, a light bracing with two cross-ties, and a heavier, more rigid layout, were compared using a free-free boundary condition to isolate the box’s own panel dynamics. Along the way, a repeatable methodology was developed for interpreting FEA mode-shape data, since raw color and amplitude values are not directly comparable across modes. That workflow was then used to distinguish brace-dominated from panel-dominated resonances and to focus design effort where it would have the most audible impact.
Fundamental panel mode frequency improved from 350Hz (unbraced) to 426Hz with added bracing.
Construction
Materials & Cutting
All panels were cut from 3/4" MDF. A full cut list was worked out ahead of time. Three 24" x 48" sheets of MDF were used to leave margin for hand-tool cutting error, practice cuts, etc.
Tools Used
Panels were cut and shaped using a circular saw with a straightedge guide, a jigsaw, a cordless drill, clamps, a combination/speed square, and a manual fixed handheld router fitted with a straight bit and an adjustable circle-cutting jig for driver cutouts.
Driver Mounting
Driver cutouts were made using a two-step technique for flush mounting: a shallow outer recess was cut first, then a full-depth inner through-hole was cut from the same center point, using the adjustable circle-cutting jig to keep both cuts concentric.
Assembly & Sealing
Hand-tool cutting produced approximately ±1–2 mm variance across panels. This will be addressed by sanding paired panels to match each other, dry-fitting the full box before gluing, assembling incrementally with square checks between joints, and sealing internal seams with wood filler.
What’s Next
All enclosure panel cutting is complete, here’s what’s left before the build is finished:
- Assemble the enclosures using wood glue and clamps.Â
- Solder the speaker wires to drivers and connection terminals mounted on the back panel.
- Leave the back panel removable (not glued or sealed yet) so the enclosure can be opened up for testing and iteration. This allows for internal adjustments to port length based on initial testing.
- Measure each driver using a UMIK calibrated measurement microphone and Room EQ Wizard (REW), running gated frequency sweep tests to get accurate quasi-anechoic response data
- Import the measured response and impedance data into VituixCAD to handle half-space response analysis and start the crossover design.Â
- Iterate on port length based on the measurements.Â
- Once frequency response, crossover point, and port length are all finalized from testing, permanently glue and seal the back panel.Â
- Enjoy good music