Three different labs across different years all slammed into the same hidden wall.
Purifi’s 1ET9040BA amplifier module created a rare problem for the lab testing it. The module was so clean that audioXpress struggled to separate its distortion from the noise inside an Audio Precision APx555B, one of the best audio analyzers available.
That would already be unusual for a high-end amplifier. But the twist is that this same module now appears in finished Buckeye monoblocks that sell for $1,295 each. A consumer amp based on it costs a fraction of the test gear that struggled to measure it.
The result says a lot about how far amplifier design has moved, and where measurement gear is starting to hit its limits.
Better Equipment, Worse Results
Axiometrix Solutions, Audio Precision’s parent corporation, supplied a second APx555B for the job.
Standard bench cables gave way to Audio Precision’s own XLR cables with Neutrik connectors and Mogami Neglex 2534 wire after the lab discovered that cable quality alone was corrupting readings.
Then, an AP AUX-0025 passive low-pass filter and Bruno Putzeys’ measurement advice completed the upgraded rig.
But while the cleaner setup did produce cleaner data, it also revealed the central problem.
At 1W, 10W, and 100W into both 4-ohm and 8-ohm loads, harmonic distortion measured below 1 part per million, or -120 dB, and in many cases significantly lower.
“One part per million” means the unwanted distortion was one-millionth of the original signal.
At that level, the analyzer is like a microphone trying to record a whisper while producing a faint hum of its own. And once the amplifier falls below that hum, the test starts describing the instrument instead of the amplifier.
That’s why the lowest-power results were the hardest to interpret. Below roughly 10 watts, the range where many speakers spend much of their time at home, the amplifier’s distortion no longer rose clearly above the APx555B’s own residual noise.
Still, this higher power gave the analyzer more to work with. As output increased, the larger signal rose farther above the analyzer’s noise floor, allowing the instrument to separate more of the amplifier from itself.
Other Labs, Same Wall
The measurement problem did not begin with the 1ET9040BA. Years earlier, other ultra-clean amplifiers had already pushed the APx555B close to the same limit.
In September 2020, SoundStage’s Doug Schneider tested a complete Eigentakt stereo amplifier built around Purifi’s first-generation 1ET400A modules.
At 1 kHz and 100W into 4 ohms, THD+N measured 0.000184%. So even with the analyzer’s full resolution applied, Schneider suspected the real figure might be lower than the result on his screen.
Benchmark’s AHB2, a Class AB amplifier, reached a similar boundary in John Atkinson’s Stereophile measurements. Its second and third harmonics at 1 kHz, 50W into 8 ohms, sat at -128 dB and -124 dB, levels Atkinson described as close to the residual distortion in the Audio Precision’s signal generator.
That pattern matters because the examples come from different reviewers, different years, and different amplifier designs.
For instance, Purifi’s first-generation Class D module and Benchmark’s Class AB amplifier both brought the analyzer near its own floor.
The 1ET9040BA makes that old ceiling newly relevant, though. It pushes deeper into the analyzer’s limits across more power levels and frequencies, then carries that performance into a module now appearing in finished amplifiers at consumer pricing.
Where the Analyzer Tops Out
At 1 kHz, the Purifi module’s -120 dB distortion and the APx555B’s typical -120 dB residual THD+N land at the same number.
The overlap matters because once the amplifier reaches the analyzer’s own residual, the reading can no longer cleanly separate the device under test from the test instrument.
However, across the full audio band, the practical ceiling can sit higher.
For one, Neurochrome’s measurement guide places the APx555B’s broadband loopback THD+N at about -114.5 dB, which becomes the limit for any amplifier being tested across that wider range. The Purifi module dips also into that territory at moderate power, leaving only part of its performance visible.
Audio Precision’s own specifications frame the same problem from another angle. Guaranteed residual THD+N sits at -117 dB, while typical 1 kHz performance reaches -120 dB.
A distortion figure that low equals one part per million of the original signal, a level far beneath audibility and squarely inside the analyzer’s hardest measurement zone.
This also explains why cheaper tools disappear from the story at this point. A QuantAsylum QA403 at $600 can reach roughly -110 dB THD+N in its sweet spot, which is more than enough for conventional amplifiers.
And for an amp operating near -120 dB, even the more than $40,000 APx555B runs out of clean separation.
Transparent Enough to Sound Wrong
Proving that an amplifier adds almost nothing to the signal settles the engineering question only so far. It still leaves the listening question, which is whether people actually enjoy that kind of transparency.
For Ralph Karsten, who designs tube amplifiers at Atma-Sphere, the problem starts with treating THD as the final word.
He called it “the least important aspect of distortion,” arguing that “low THD does not correlate with the most musical sound.” And, he has also conceded that “Class-D can be better” than other topologies, so his objection is narrower than a rejection of the format itself.
In short, his target is the assumption that measurement perfection automatically becomes musical satisfaction.
Meanwhile, Didden made the same debate feel less abstract in his audioXpress article. Many listeners, he observed, are so accustomed to the warmth and subtle artifacts of traditional amplifiers that they describe contemporary Class-D designs as “clinical and lifeless,” as though perfect transparency sounds like something is missing.
What they hear as missing, Didden noted, was never part of the music. It was a contribution from the amplifier.
The measurements confirm the 1ET9040BA adds almost nothing to the signal, which for some listeners is exactly the problem. But while measurements can show transparency, they cannot decide whether anyone has to prefer it.
When the Instrument Costs More
The Purifi 1ET9040BA module that audioXpress struggled to measure doesn’t sell directly to consumers. But the same module sits inside Buckeye’s finished monoblocks, which ship as complete amplifiers for $1,295 each, or $2,590 for a stereo pair.
An analyzer costing more than $40,000 couldn’t fully characterize the module. A single Buckeye monoblock runs less than a fifteenth of its price, and a pair runs less than a tenth. The most expensive component in the measurement chain turned out to be the instrument, not the amplifier, and the instrument was the one that couldn’t keep up.
Putzeys co-founded Purifi and designed the Eigentakt modules, so he has an obvious stake in the narrative. But independent measurements from other labs and other amplifier classes reached the same ceiling, giving his provocation weight it didn’t carry when Class D still had something to prove.
Instrument resolution has become the bottleneck, not amplifier quality. The engineering argument, for practical purposes, has been settled by the numbers, but the listening argument, as Karsten and a substantial share of the audience would insist, was never going to be settled by an instrument in the first place.