“NIBIRU” is the somewhat ironic name of an experimental preamplifier created to test some widely held audiophile beliefs: the supposed “magic” of even-order harmonics generated by vacuum tubes, the demonization of negative feedback, and the idea that a “zero-feedback” circuit can add three-dimensional sound that was never present in the original recording. This project, conceived as a true test bench, aims to measure, listen, and understand what really happens to an audio signal when it is subjected to different types of loading and feedback.
Why the name “Nibiru”?
Nibiru is the mythical planet invented by writer Zecharia Sitchin, entirely devoid of scientific evidence. Likewise, the world of high-fidelity audio is full of its own “imaginary planets”: dogmas, legends, and unfounded convictions that nonetheless influence design choices and marketing.
Myths to debunk: harmonics and feedback
According to a popular story, the charm of vacuum tubes lies in the fact that they generate only even-order harmonics, while push-pull circuits produce odd-order harmonics considered less pleasant. In reality, any triode generates a mix of even and odd harmonics, in proportions that vary with the circuit, the operating point, and even the individual tube. Another myth claims that negative feedback “erases sonic information.” In truth, the absence of feedback does not preserve signal purity—it adds distortion and irregularities that may sometimes please the ear, but are not present in the original signal.
Distortion: not just harmonics
Harmonic distortion is only one of many possible distortions. There are also phase distortion, intermodulation, and memory effects, often more influential on the sonic result. Showing and interpreting all these factors would require sophisticated instrumentation, but even with the available measurements the conclusion is clear: reality is far more complex than a single THD percentage on a datasheet.
The experimental circuit
Nibiru uses a single ECC82/12AU7 tube (or equivalents 6189, 5814A, 12BH7, 6CG7, 6SN7) powered at a low voltage of about 80 V. A four-pole switch provides two operating modes:
- Resistive – anode loaded with a resistor and an un-bypassed cathode resistor: practically no reactive elements.
- Reactive – cathode bypassed and anode resistor replaced by a 19S555 inductor: a classic parafeed configuration enriched with variable local negative feedback, the distinctive feature of this experiment.
Two dual potentiometers complete the circuit:
- a tone control to adjust low frequencies;
- a continuously variable negative feedback control.
Listening test
The experiment consists of inserting Nibiru between a known high-quality source (for example a DAC or CD player) and a well-feedbacked power amplifier with adequate damping of the loudspeakers. A quality amplifier is preferable: a poor or zero-feedback unit would distort the results.
At this point you can experiment freely, adjusting the potentiometers to hear how the sound changes. The experience is instructive and sheds light on phenomena that usually remain purely theoretical.
Resistive mode
In the resistive configuration, free of reactive elements, Nibiru behaves almost neutrally. The only audible effect is a slight pre-amplification and the minimal harmonic distortion typical of the tube. With maximum negative feedback even that small distortion decreases, making the difference from a direct connection almost imperceptible.
Reactive mode
The character of the circuit changes dramatically when the inductor is engaged, creating a true parafeed stage. Nibiru adds an original idea: variable local negative feedback.
Listening reveals a very distinct sequence of changes:
- Maximum feedback – sound is rather closed-in and stuck to the speakers. Curiously, some listeners prefer it this way, confirming how subjective “better or worse” really is.
- Gradually reducing feedback – the soundstage opens, detaches from the speakers, a central front forms, and instruments and voices distribute naturally across it.
- Mid feedback (around 1 o’clock on the pot) – a slight blending of left and right channels begins.
- Minimum feedback – the mid-high range remains airy, but low frequencies lose definition, becoming first a bit muddy, then increasingly bloated and intrusive. The tone control can tame the level of bass, but not fully restore clarity.
These observations refer to a load given by an amplifier’s input potentiometer; with a direct loudspeaker load, interactions become even more complex.
Measurements
Tests on resistive and reactive loads confirm these listening impressions:
- Total harmonic distortion (THD): ranges from 0.4 % to 1.3 % depending on mode and feedback.
- In all cases there is a mix of even and odd harmonics (2nd, 3rd, 5th, 7th, 9th), disproving the idea that tubes generate only even harmonics.
- Frequency and phase response: the reactive load introduces greater phase rotation.
- 10 kHz square-wave response: waveforms show rounding and ringing that vary with feedback level.
Spectrum analysis, while detailed, does not offer definitive explanations: the small fractional-percent THD differences between configurations do not account for the striking audible changes. It is also worth noting that even a simple single-triode stage never produces only even harmonics, with or without feedback.
The spectrum without feedback shows harmonics 2nd, 3rd, 5th, and 9th, while with maximum feedback it shows 2nd, 3rd, 7th, and 9th. The reactive mode presents a similar harmonic mixture. Anyone claiming that tubes generate exclusively even harmonics is not describing reality.
Some might object: “But you can hear a 0.1 % difference!” Experience proves otherwise. Replacing the ECC82 with other ECC82s, or compatible types such as 6211, 5814, or 12BH7, produced THD changes larger than 0.1 %, yet no audible difference. Blind tests over the years confirm that THD differences below about 0.5 % are detectable only by instruments, not by the human ear.
With these results in hand, it is natural to proceed to an analysis of the frequency and phase response, where further aspects of the circuit’s behaviour become evident.
Because of a scale setting error, phase plots are not directly comparable: 50 degrees per division for the resistive load and 10 degrees per division for the reactive load. Even so, the phase behaviour of the resistive load is clearly superior, with about 25 degrees of rotation starting at 1 kHz. This alone shows that reactive elements cause greater phase shift. A 10 kHz square-wave test further illustrates the point, where the generator’s signal is shown in yellow and the circuit’s response in blue.
On a resistive load nothing noteworthy appears apart from amplitude change. Rise and fall edges are only slightly rounded, due to the circuit’s stray capacitances. With an inductive load and no feedback the square wave is more rounded and distorted. As feedback is increased the deformation gradually decreases, but never disappears. At maximum feedback small ripples appear on the wave tops—ringing caused by a minor internal resonance.
These behaviours show that the most significant effects occur on transient signals, where capacitors and inductors inject additional components. Negative feedback can partially cancel these additions (which were introduced by the circuit itself, not contained in the source signal) and, through interaction, can also create new ones. More sophisticated instrumentation could describe these effects more precisely, but the tendencies are already clear.
Final considerations
In conclusion, it is clear that reactive elements add distortion to the signal, distortion unrelated to simple harmonics. This means that many zero-feedback circuits and the use of interstage transformers are aimed above all at coloring the sound, adding features that were not originally present. It is not negative feedback that “erases” information; it is the lack of feedback that generates new components. These can sometimes be pleasant, but—as demonstrated by the Nibiru experiment—they also bring unwanted side-effects such as sonic smearing and loss of definition.
Experience suggests that the best listening result comes from finding a balance: carefully adjusting the amount of negative feedback until you reach the sweet spot that provides a broad, engaging soundstage without the haze and artifacts typical of total feedback absence. Anyone wishing to repeat the experiment more simply can build only the reactive section with adjustable feedback, omitting the selector and tone control.
Nibiru shows that many audiophile “truths” are really myths. A touch of well-judged negative feedback works like a perfectly balanced seasoning: it enhances the music’s flavour without covering it.
