Methodology: How we research and verify

PASpeaker-Base does not aim to collect the most build plans, but the most reliable ones. That is why every technical figure has a provenance, a source and a verification status. Anything that is not documented does not appear here.

1. Invent nothing. If a value is missing from the source, the page says “not available” – never an unlabelled estimate.
2. Keep measurement and simulation apart. Simulated curves are not measurements and are always labelled as simulations.
3. The community does not replace manufacturer data. We show experience reports from forums in addition, clearly labelled.
4. No substitute drivers by size. We only name a different driver with a justification: approval by the designer or manufacturer, a simulation or a measurement.
5. Link to third-party plans, do not copy them. We do not host third-party drawings, graphs and PDFs; we link to the original source.

Next to every parameter it says where it comes from: from the manufacturer, from the designer, measured, simulated, calculated by us, from the community or estimated. Values from drawings or tables also state where they were found, for example “PDF p. 6”.

Measurements come with their conditions: distance, normalisation (e.g. 2.83 V / 1 m), placement and smoothing. If any of this is missing, we label the measurement as only partly comparable.

• Verified: All core data has been checked against the primary source.
• Partially verified: The primary source has been checked; individual figures are missing from it or still open.
• Community design: There is no manufacturer or designer data on performance. The information comes from Tier B and C sources.
• Unverified: Not recommended and not listed in search engines.

Sometimes figures contradict each other – even within a single manufacturer document. In that case we show both values with where they were found and explain which one we adopt. For geometry, the drawing usually takes precedence, because that is what you build from. We never take an average.

We automatically check driver data for internal consistency, for example whether Qts matches Qes and Qms, or Vd matches Sd and Xmax. The result is shown openly on the respective driver page.

The enclosure designer calculates with the Thiele and Small model. It treats horns, tapped horns and transmission lines as a chain of short pipe sections. We check it in four ways, each with automated tests:

• Against theory: Sealed enclosures, bass reflex alignments according to Thiele, the quarter-wave resonance of a line, the radiation impedance and the group delay must match the known formulas.
• Against the datasheets: The sensitivity of a driver can be calculated from the Thiele-Small parameters. For 75 drivers in our database, the result has a median deviation of 0.1 dB from the manufacturer’s figure and is never more than 2.7 dB away. This is within the usual datasheet tolerances.
• Against a measurement: For the Celestion reference design CKT-FM-FTX1225 (48 L, 48 Hz), the designer calculates a lower cut-off frequency of 64 Hz (−3 dB) and 40 Hz (−10 dB). Celestion measures 65 and 44 Hz on the floor.
• Against its own build plan: A stress test builds every driver in the database in every enclosure type, single and double, and checks the plans: panels inside the enclosure and without overlaps, net volume accurate to 2 %, ports inside their chamber, tubes – including those with a bend – clear of panels and driver, braces and struts not in the driver, the driver including its magnet does not hit any panel, all files complete. On 14 November 2026 this revealed that too little of the air column in tube ports was subtracted from the enclosure volume. As a result, plans with tubes were too small, by around 13 L for 2 × Ø 150 mm in 200 L. In addition, some sealed enclosures and most rear chambers of front-loaded horns were shallower than the driver is deep. Both errors have been fixed: the enclosure is made at least as deep as the driver, and where the volume is too small for that, a filler block compensates. For the tapped horn and the TL (transmission line), the build plan names the cases in which the line is too shallow for the driver.

Simulations remain estimates. Under large-signal operation, with room influence and with a differently built enclosure, real values deviate. That is why we label every result of the designer as calculated.

Every page states when it was last checked. We check links to original documents regularly. For every document we store a checksum so that we can detect silent changes.

Reports of errors or better sources are expressly welcome. The contact details are in the legal notice.