I updated this article because I want to better express my thoughts on a topic that’s hotly debated on the internet: the use of negative feedback in audio amplifiers. I’m tired of sterile arguments and attacks on those who don’t conform to common thinking. Here I say my piece, without schematics, graphs, or formulas: I prefer to speak from practical experience and debunk some myths that fuel a real witch hunt.

Negative feedback is an important concept in audio electronics with roots in telephone system theory. In short, feedback is a mechanism used to reduce distortion and improve the stability of a system. Feedback was first used in audio electronics in the 1930s, when Harold Black, an engineer at Bell Laboratories, developed a feedback amplifier to improve the performance of the triode amplifiers then in use. Black’s feedback amplifier used a negative feedback loop to reduce amplifier distortion. Negative feedback made it possible to significantly lower background noise and harmonic distortion and to increase the bandwidth of an amplifier circuit, allowing longer lines that could reach many more users, with a quality of telephone transmission otherwise impossible. The invention of negative feedback therefore gave technology a huge boost.

In the 1940s and 1950s the concept of feedback was applied to various electronic devices, including amplifiers and mixers. Over the years feedback evolved, becoming a fundamental part of audio circuit design. The principle is simple: part of the output signal is fed back to the input and compared with the original signal. Any difference (error) is subtracted, so the amplifier corrects its imperfections in real time. The result is a response that’s more faithful to the original signal, with reduced distortion and noise and a cleaner sound quality. Today feedback is used in many audio applications (amplifiers, mixers, equalizers, and compressors) and non-audio ones (automotive, industrial control). In short, feedback is a concept that has improved the quality and stability of audio systems and beyond.

To read something more technical about feedback I recommend this article on Audio Valvole.

Zero-Feedback Tube Amplifiers

What are the arguments of zero-feedback supporters? Supporters of zero-feedback audio amplifiers claim these amplifiers produce a sound that is more “warm,” “faithful,” and “more natural” than feedback-equipped amplifiers. In particular, they argue that tube amplifiers produce harmonic distortion that’s more pleasing to the human ear than the distortion caused by heavily feedbacked amplifiers, which is generally considered more “harsh.”

They also believe these amplifiers can handle signal peaks better, ensuring a more dynamic and “realistic” reproduction. It is claimed that tube technology can handle peaks without causing audible compression.

What is Clipping?

Clipping in an audio amplifier occurs when the output signal exceeds the amplifier’s maximum handling capability. When the input is too strong and the amplifier cannot further increase the voltage without distorting, the waveform is “cut” at its maximum limits, causing distortion. It’s heard as a “squashed,” “messy,” or “raspy” sound, depending on severity.

In tube amps, clipping tends to appear more gradually, mainly due to the transfer curve of the tubes and their biasing: the amplification stops increasing progressively and the peaks round off. The output transformer can contribute, for example by limiting the bandwidth or by saturating in extreme cases, but it is not the primary cause of this effect. This kind of clipping is not very bothersome to the ear or is almost unnoticeable if it happens for brief instants.

In solid-state amplifiers, by contrast, clipping often appears more abruptly and sharply: once the voltage limit is reached, the waveform is cut suddenly, producing a clear and unpleasant distortion.

It’s important to clarify that soft clipping does not mean greater dynamics: if peaks exceed the available power they are still cut. To truly reproduce a track’s dynamics you need to lower the volume or use a more powerful amplifier. Anyone claiming that a tube circuit can, even for an instant, deliver power much higher than its nominal rating is telling astronomical lies: this is, if anything, a characteristic of certain solid-state stages (because of reserves in capacitors or particular configurations). A tube amp, when it clips, clips, period.

Zero-Feedback Myths

Zero-feedback amplifiers—i.e., those that do not use negative feedback—are often the subject of myths and misunderstandings. Below are some of these myths and the related corrections:

• “Zero-feedback amplifiers sound better than feedback amplifiers”: that’s a generalization. When designed and applied correctly, negative feedback can reduce distortion and improve linearity, ensuring a sound that is faithful to the source.
• “Zero-feedback amplifiers are more musical and warm”: it depends on the design, not just on the presence or absence of NFB. There are many feedbacked tube amps with a very “musical” sound.
• “Zero-feedback amplifiers have greater dynamic range”: not automatically. Perceived dynamics depends on the whole chain and on real voltage and current limits.
• “Zero-feedback amplifiers are more reliable”: reliability depends on design and components, not solely on the presence or absence of NFB. In fact, a circuit without feedback tends to require more frequent maintenance, because even small variations in components—especially tubes—can cause differences between the two channels of the same device in relatively short times.
• “Feedback introduces distortion or noise”: within the band where loop gain is sufficient, NFB generally reduces distortion and noise. Problems can arise only in poorly compensated designs or with insufficient phase margin.

Limitations and Negative Effects of No Feedback

1. Higher distortion: without NFB the distortion remains the circuit’s “native” one and can be higher (especially at the band edges where the system is less linear).
2. Higher noise: without the NFB “safety net,” perceived noise can be higher (it also depends on the power supply).
3. Higher output impedance: limits the ability to control the load (low damping).
4. Greater sensitivity to tolerances and variations: behavior depends more on components, speakers, and environment.

In general, the absence of feedback in a tube amplifier can have negative effects on sound quality. However, in some configurations—e.g., with horn speakers or single-driver systems featuring small, high-efficiency drivers—it can be pleasing to those who prefer a certain kind of timbre. Conversely, with low-impedance speakers, complex crossovers, large woofers, or reflex loads, the lack of feedback almost always leads to a disorderly sound: boomy bass, poor precision, and difficulty reproducing complex musical passages, which end up turning into confusion.

Feedbacked Tube Amplifiers—But in the Right Amount!

Using a modest amount of negative feedback in a tube amplifier can offer several advantages:

1. Reduced harmonic distortion: improves sound fidelity.
2. Improved frequency response: more even amplification across the useful band.
3. Lower background noise.
4. Higher damping factor and thus better woofer control.

In summary, moderate feedback can improve stability, linearity, frequency response, and noise. The amount and type of feedback must be balanced with the desired sonic characteristics.

If There’s Too Much Feedback?

An amplifier with too much feedback or with wrong compensation can yield:

1. Sound perceived as “too dry” with certain pairings (very high damping factor, different interaction with the speaker/room).
2. Perception of “sterile” in poorly balanced designs (this is not automatic with NFB).
3. Stability issues (self-oscillation) if the phase margin is insufficient.

Feedback can affect the signal’s phase rotation: with high loop gains at certain frequencies, conditions for instability can be created. Moreover, poor management of phase and loop can increase intermodulation. These problems are reduced with careful design and tuning. In a quality tube amp the bandwidth of the output transformer must be as wide as possible.

Feedback and Damping Factor

Another widespread belief is that a very high damping factor reduces a system’s “dynamics.” In reality, a high damping factor mainly indicates a low output impedance, which allows the amplifier to better control woofer motion and reduce unwanted resonances.

It may happen that, on certain speakers or in certain rooms, a very controlled bass is perceived as more “dry” or “less full-bodied,” but this is not a real loss of dynamics: it’s a different interaction between amplifier, speakers, and room acoustics. For more: dedicated article.

What’s the Definition of Hi-Fi?

Hi-Fi means “High Fidelity”: reproducing the source signal as much as possible, with detail and balance, reducing colorations and distortions. The goal is to offer a reproduction that respects what was recorded.

Criticalities in Feedbacked Circuits

Feedback can lead to instability and self-oscillation if the circuit is not properly designed or is poorly built. Criticalities to consider:

1. Phase: if rotation exceeds 180° with gain still high, the loop takes off.
2. Bandwidth: the loop introduces delays; phase margins and compensation are crucial.
3. Gain: too much loop gain without adequate margins = oscillations.
4. Impedances: impedance interaction can affect stability.
5. Timing: every loop needs time to act; careful design is required.

To avoid instability you need careful design, component choices, correct configuration, and analysis of stability parameters. A zero-feedback design can be simpler to build without oscillations, which is why in the early days of DIY such schemes were often proposed; then it also became a “trend.”

Is the Absence of Feedback Necessary for Natural Listening?

No. “Natural” listening depends mainly on the overall quality of reproduction (chain, acoustics, recording), not on the loop type alone. The absence of NFB can imply higher distortion, greater sensitivity to variations, and more noise. There are feedback configurations that improve fidelity without compromising pleasantness.

Artisanal Amplifiers VS Industrial Amplifiers

Quality depends on design, components, construction, and care. Industrial builds can benefit from repeatable processes; artisanal builds from customization and selected components. At the industrial level more NFB is often used for uniformity and specs; at the artisanal level it’s chosen according to the desired sonic result. Watch out for the unpresentable (examples here)…

Down with Extremists

There are supporters of “absolute zero feedback” who believe that using feedback negatively affects sound. It’s a legitimate preference. Some, however, become dogmatic and critical of those who use it in a moderate and informed way. In reality, using feedback can be useful and beneficial if well implemented. The amount can be adjusted to taste and system. It should not be treated dogmatically.

It’s also worth considering that when a track is recorded, between preamps, mixers, recorders and everything else, the signal passes through a large number of professional devices full of feedback. If NFB really “ruined” the sound, why should the last bit of feedback in your home system be the culprit? Final distortion often occurs in playback, especially if you use a “natural/zero-feedback” amplifier outside its ideal context.

Saying “a feedback amplifier sounds bad” makes no sense: it depends on the circuit, components, design, and type of feedback. Properly used NFB leads to precise, faithful reproduction. You can’t generalize. There are thousands of tube amps with NFB appreciated by lots of people. Tastes matter.

Zero-feedback fundamentalism can alienate potential tube-amp users by making this world seem dominated by a rigid ideology. Moreover, disproportionate criticism of feedback amplifiers can discourage those seeking performance achievable realistically only with a certain degree of feedback.

My experience: I’ve met people convinced by “zero-feedback” theories who spent large sums on amplifiers built according to this philosophy, only to be disappointed. In many cases there was a lack of practical electronics knowledge. The result was a slow sound, confused on complex passages, with bloated bass and vagueness. When they were able to listen to a properly feedbacked and well-designed amplifier, they perceived greater width, dynamics, depth, and cleanliness. Buying an amplifier should be informed and conscious, not dictated by fashion or dogma.

When It Becomes a Sickness…

Over the years, stories like the following have happened to me at least 3–4 times… Caius, a client of mine, some time ago bought from me a set of transformers and a premium schematic to build one of my KT88 amplifiers, a precursor to Phoenix. One day he meets “Titius,” a friend who’s a tube-amp enthusiast. Titius has a zero-feedback 300B built by “name of super-guru”, but he’s not convinced by the sound. Caius listens: mids-highs OK, in the lows rubbery and slow. Then Titius listens to Caius’s KT88: “Wow it sounds good, I wish mine sounded like that too!” — “Have SB-LAB build you the same.” — “Ehhh but you know the coooooontrooooooreaaaaactiooonnnn…”.

The next day Titius buys a pair of vintage 300B for €2500 for his zero-feedback amp. The following week he puts everything up for sale (amp + 300B). The following month he buys another zero-feedback 300B by “another super-guru”. He has already thrown away almost €20,000 and the story repeats. There are many “Titiuses” like this…

How I Conceive Tube Amplifiers… My Conclusions

Tube amplifiers have a unique charm. The choice between zero feedback and feedback depends on goals and context. Personally I aim for a balance between bass management and mid-high clarity, without becoming hard and aggressive like certain solid-state designs nor clumsy and slow like some zero-feedback ones.

Weighted use of feedback: it’s an important tool to achieve a more linear response, lower distortion, and better control of demanding speakers. One advantage is versatility with more speaker types. Zero-feedback supporters often say you need “dedicated” speakers like horns or single-drivers: this restriction complicates life for those who want to freely choose their system.

Obviously NFB requires a solid design: a capable driver stage, a transformer with a wide bandwidth, a clean power supply, and adequate phase margins. With this approach you get a balanced, repeatable sound on real-world systems.

I want to make it clear that I fully understand the techniques behind zero feedback and could apply them. My choice to use moderate feedback comes from experience and from the idea of sound I seek. Zero feedback is not a panacea and brings concrete limits; feedback is not “evil.” It’s a tool: it depends on how you use it. In short: no extremism. The choice depends on many factors. Amplifier design is an art of balance and common sense more than slogans.