So what is deep negative feedback and how is it influencing sound?

[DNCAgain]

I recently came across a viewpoint on a video website that companies like Topping use "deep negative feedback" in their products, which makes their THD (Total Harmonic Distortion) data look more impressive but leads to a "smearing" effect (loss of details) and a loss of dynamics.

However, given that he once had a dispute with Topping, it's hard to judge the accuracy of this claim.

Below is another discussion from the same video website about amplifier.


“Since common integrated operational amplifiers have a very narrow open-loop bandwidth (a few hertz to several kilohertz), they are often used in amplifiers with deep loop negative feedback applied to extend the passband. However, this approach inevitably introduces severe transient intermodulation distortion (TIM), degrading the sound quality. Therefore, amplifiers built with integrated op-amps, while capable of achieving good steady-state performance through deep negative feedback, cannot be considered high-fidelity (Hi-Fi) from a transient perspective.”

 

[DVDdoug]

but leads to a "smearing" effect (loss of details) and a loss of dynamics.

TOTAL NONSENSE!!!

Negative feedback (corrective feedback) has all kinds of wonderful effects. Lower noise, lower distortion, and flatter frequency response. With an op-amp (feedback required) you can make perfect-sounding audio circuit, almost without thinking.

You can't drive a car or navigate an airplane or ship without feedback. You have to know where you are and what direction you're going, and in a car it's especially important to monitor, and correct, speed.

I have no idea what "deep" feedback is...

 

[DNCAgain]

TOTAL NONSENSE!!!

Negative feedback (corrective feedback) has all kinds of wonderful effects. Lower noise, lower distortion, and flatter frequency response. With an op-amp (feedback required) you can make perfect-sounding audio circuit, almost without thinking.

You can't drive a car or navigate an airplane or ship without feedback. You have to know where you are and what direction you're going, and in a car it's especially important to monitor, and correct, speed.

I have no idea what "deep" feedback is...

What about TIM?negative feedback has a positive or negative impact on it?

 

[TubeDesigner]

TIM was cured as an issue decades ago. As far as "deep feedback", I think their marketing department has been a bit too "creative", as most audio companies marketing departments are. Move on, nothing to see here.

 

[oleg87]

leads to a "smearing" effect (loss of details) and a loss of dynamics.


“Since common integrated operational amplifiers have a very narrow open-loop bandwidth (a few hertz to several kilohertz), they are often used in amplifiers with deep loop negative feedback applied to extend the passband. However, this approach inevitably introduces severe transient intermodulation distortion (TIM), degrading the sound quality. Therefore, amplifiers built with integrated op-amps, while capable of achieving good steady-state performance through deep negative feedback, cannot be considered high-fidelity (Hi-Fi) from a transient perspective.”

Complete and utter nonsense.
Any time an audiophile starts talking about negative feedback you can pretty much assume they are clueless or trying to sell you something.

 

[DNCAgain]

That’s so true! Soon after, he promoted a purely analog-designed amplifier, claiming its sound was better than that of amplifiers with more scientific designs.

Complete and utter nonsense.
Any time an audiophile starts talking about negative feedback you can pretty much assume they are clueless or trying to sell you something.

 

[oleg87]

TIM was cured as an issue decades ago. As far as "deep feedback", I think their marketing department has been a bit too "creative", as most audio companies marketing departments are. Move on, nothing to see here.

I think Topping uses the term NFCA to refer to their higher-order feedback topologies. "Deep feedback" is some term the writer came up with.

The obligatory Putzeys writeup on the topic, from a true expert in the field: https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf

 

[DNCAgain]

Thanks for sharing!

I think Topping uses the term NFCA to refer to their higher-order feedback topologies. "Deep feedback" is some term the writer came up with.

The obligatory Putzeys writeup on the topic, from a true expert in the field: https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf

 

[JiiPee]

I recently came across a viewpoint on a video website that companies like Topping use "deep negative feedback" in their products, which makes their THD (Total Harmonic Distortion) data look more impressive but leads to a "smearing" effect (loss of details) and a loss of dynamics.

However, given that he once had a dispute with Topping, it's hard to judge the accuracy of this claim.

Below is another discussion from the same video website about amplifier.


“Since common integrated operational amplifiers have a very narrow open-loop bandwidth (a few hertz to several kilohertz), they are often used in amplifiers with deep loop negative feedback applied to extend the passband. However, this approach inevitably introduces severe transient intermodulation distortion (TIM), degrading the sound quality. Therefore, amplifiers built with integrated op-amps, while capable of achieving good steady-state performance through deep negative feedback, cannot be considered high-fidelity (Hi-Fi) from a transient perspective.”

Claims of the audiophool salesmen typically fall in one of the following categories:
1. The claimed phenomenon is totally nonexistent and against the laws of natural sciences.
2. The claimed phenomenon does exist, but the effect is so small that it is well below the human hearing capabilities, and therefore meaningless in audio context.
3. The claimed phenomenon does exist, but only in frequencies that are way above the human hearing capabilities, and therefore meaningless in audio context.

Claims that "Negative feedback create transient response problems" fall into category 3.

 

[DNCAgain]

Claims of the audiophool salesmen typically fall in one of the following categories:
1. The claimed phenomenon is totally nonexistent and against the laws of natural sciences.
2. The claimed phenomenon does exist, but the effect is so small that it is well below the human hearing capabilities, and therefore meaningless in audio context.
3. The claimed phenomenon does exist, but only in frequencies that are way above the human hearing capabilities, and therefore meaningless in audio context.

Claims that "Negative feedback create transient response problems" fall into category 3.

Specifically, at what frequency does it affect sound? For example, 1 MHz?

 

[NTK]

Specifically, at what frequency does it affect sound? For example, 1 MHz?

It depends on the amplifier design.

Below is from an article by TI on OpAmps. The blue curve is the OpAmp's open loop gain with signal frequency (example is OPA211, the article is for general opamp applications and not audio specific). The open loop gain curve is specific to the amplifier, and each design will have its own curve. You can see that open loop gain drops with frequency (the amplifier displays a low pass characteristic) and it is typical to most amplifiers.

The green curve is the closed loop gain curve, and this curve shows how this amplifier is designed to be used in this example case. In this example the amplifier has a designed closed loop gain of +46 dB. The difference between the blue curve and the green curve is the amount of negative feedback (you can look at it as the "excess" gain getting "consumed" by the feedback). Because the open loop gain drops with frequency, we have progressively less feedback as frequency goes up. That's the primary reason why harmonic distortions goes up with frequency -- it is because we have less and less feedback available. The intended usable frequency range of this amplifier will end somewhere before the blue curve meets the green curve, depending on the application requirements.

 

[DNCAgain]

It depends on the amplifier design.

Below is from an article by TI on OpAmps. The blue curve is the OpAmp's open loop gain with signal frequency (example is OPA211, the article is for general opamp applications and not audio specific). The open loop gain curve is specific to the amplifier, and each design will have its own curve. You can see that open loop gain drops with frequency (the amplifier displays a low pass characteristic) and it is typical to most amplifiers.

The green curve is the closed loop gain curve, and this curve shows how this amplifier is designed to be used in this example case. In this example the amplifier has a designed closed loop gain of +46 dB. The difference between the blue curve and the green curve is the amount of negative feedback (you can look at it as the "excess" gain getting "consumed" by the feedback). Because the open loop gain drops with frequency, we have progressively less feedback as frequency goes up. That's the primary reason why harmonic distortions goes up with frequency -- it is because we have less and less feedback available. The intended usable frequency range of this amplifier will end somewhere before the blue curve meets the green curve, depending on the application requirements.

View attachment 521039

That's very clear!Thanks

 

[Mnyb]

There is always a myth in hifi that negative feedback is bad , it comes ups now and then .

The truth badly designed amps are bad sometimes badly designed feedback can be involved ?

 

[Chris A]

I recently came across a viewpoint on a video website that companies like Topping use "deep negative feedback" in their products, which makes their THD (Total Harmonic Distortion) data look more impressive but leads to a "smearing" effect (loss of details) and a loss of dynamics.

Moderate amounts of negative feedback result in increased higher order harmonic distortion, which makes the sound quality more opaque:

This from Nelson Pass.

Using multiple gain stages in amplifiers with moderate amounts of negative feedback results in increased modulation distortion:

Read the linked article above to understand these plots more fully. Nelson Pass wrote an excellent article on the subject.

Chris

 

[solderdude]

Moderate amounts of negative feedback result in increased higher order harmonic distortion, which makes the sound quality more opaque:

With harmonic distortion also comes higher amounts of IM distortion which is NOT harmonically related nor 'pleasant'.
You cannot create a non linear distortion that does not also create IM products.
Whatever Nelson Pass writes is to increase sales (advertisement) from his amps.

It is a myth that tiny amounts of mainly lower harmonics has a 'beneficial effect' on perceived sound quality.
That would need to be proven in a well executed blind test.
Below audible thresholds is and always will be below audible thresholds.

 

[Chris A]

Whatever Nelson Pass writes is to increase sales (advertisement) from his amps.

I think you went a bit too far with this statement (...to put it very mildly...).

Go over to diyAudio (a huge forum) and read a bit there to better understand what his contributions are to amplifier design. It's not to sell more Pass Labs or First Watt amplifiers--just the opposite.

Chris

 

[Chris A]

It is a myth that tiny amounts of mainly lower harmonics has a 'beneficial effect' on perceived sound quality.

I agree with this.

I think what Mr. Pass is saying, however, is if you have real devices with nonlinear gain curves (which includes all devices used for amplifiers), its better to trade for lower order harmonics than higher order. Also, it's usually better to trade for fewer gain stages in amplifiers than more--to keep the modulation (IM) distortion down as much as possible.

Chris

 

[solderdude]

So ... you are telling me Nelson Pass is putting out all this info to make sure no one will buy his designs but should buy anything that has lower distortion than his (purposely higher distortion) designs.

I am not stating he does not know anything about amplifier design, he does know a lot, like many other designers do as well, and he can design super low distortion amps as well but chooses to sell higher distortion amps that all will also have IM distortion.

Nelson Pass is in the business of selling amps, just like all other manufacturers.
He is very vocal about his design philosophy, so are other designers and they all sell something that 'sounds better' than other brands, they all have a (cult like) following and people that swear by designs made by this or that guru.

And that's exactly what Nelson Pass is to many people ... a guru.

 

[solderdude]

I agree with this.

I think what Mr. Pass is saying, however, is if you have real devices with nonlinear gain curves (which includes all devices used for amplifiers), its better to trade for lower order harmonics than higher order. Also, it's usually better to trade for fewer gain stages in amplifiers than more--to keep the modulation (IM) distortion down as much as possible.

Chris

But that is just his opinion which is reflected in his designs.

Amps that simply have low distortion (copious amounts of open-loop gain and feedback) arguably have low harmonic and low IM distortion.
Arguably this is bandwidth related as well as we can see in low HD class-D amplifiers that have substantially lower THD in higher frequencies and have high amounts of IM in the higher frequencies.
This is easily seen in multitone plots of class D amps (you see the typical rise in the higher frequencies).
Fortunately humans are not as sensitive to both lowest and higher frequencies so it is not detrimental to the sound.

 

[Chris A]

So ... you are telling me Nelson Pass is putting out all this info to make sure no one will buy his designs but should buy anything that has lower distortion than his (purposely higher distortion) designs.

I think this sentence reflects a little decision bias on its own, to be honest. In other words, I don't really think it's a fair assessment of the situation. If you actually read his articles and posts, I think the situation is a lot different than you imply.

That's okay to trade differently than Mr. Pass does, but I think your comments are more than just a bit harsh here. Given his contributions to, for instance, yours, to hi-fi amplifier practice, I think I'd still pick Mr. Pass's designs. But that's just me.

It's just audio...

Chris