• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

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

DNCAgain said:
My statement was incorrect, I apologize.

In fact, his description is completely the opposite. The advantages he claimed are as follows:

1. Pure digital design (no need for digital-to-analog conversion);
2. Feedback-free design;
3. GaN material (provides greater thrust);
4. PCM to PWM framework.

However, he did not show any data except for labeling the power as 500W×2!
Click to expand...
Thrust, eh?
hmm...
units of thrust (force): Newtons (N): One N = 1 (kg)(m)/sec^2
power = force x velocity
I mean, I guess amplifiers could be spec'd based on thrust...
My brain hurts.
 
These kinds of designs exist, but I’m sure most wil use some kind of feedback, or have performance issues.
Click to expand...
Actually, no. I've done some projects with a microcontroller that has PWM built-in (or it can be done in software). You can feed-in a digital value (i.e. an audio sample) and output a related/proportional PWM value Which can be filtered to an analog voltage. It's all software computation and no digital or analog feedback is required.

This particular microcontroller has an ADC but no DAC. So you could have an analog input, and still, no feedback would be needed. (In reality, the particular Arduino microcontroller that I've used only has a 10-bit ADC and 8-bit PWM, and it's slow. So, it's NOT for high quality audio.)
 
DVDdoug said:
Actually, no. I've done some projects with a microcontroller that has PWM built-in (or it can be done in software). You can feed-in a digital value (i.e. an audio sample) and output a related/proportional PWM value Which can be filtered to an analog voltage. It's all software computation and no digital or analog feedback is required.
Click to expand...
Yes, you can do PWM without feedback. That doesn’t mean you can make a well performing amp out of it, though. Not the same thing!
 
mhardy6647 said:
Thrust, eh?
hmm...
units of thrust (force): Newtons (N): One N = 1 (kg)(m)/sec^2
power = force x velocity
I mean, I guess amplifiers could be spec'd based on thrust...
My brain hurts.
Click to expand...
At the power amplifier company where I used to work, we rated one of our amplifiers in horsepower. Seriously - sounds like something Cerwin-Vega would do, but there you are! ;)
 
TIM is actually caused by having too little feedback (ie too little loop gain) in combination with an input stage that overloads because of the reduced loop gain.
 
MaxwellsEq said:
To answer your questions correctly, you need to have a bit of the basics of control theory. At first glance, it's all quite obvious. Once you get into it a bit deeper, the mathematics can seem a bit heavy, but is essential for designing stable systems when the bounds are challenging (e.g. when rates of change are large).

There's nothing in the theory that says control systems are bad. Quite the opposite, since a system without controls may have runaway characteristics leading to damage to the system or the person operating it! However, the control system needs to have the same level of requisite complexity as the system it controls.
Click to expand...

From the article I shared yesterday The Age of the Amplifier by Brian Potter.

By the end of the year [1927], [Harold] Black had built a negative feedback amplifier that reduced distortion by a factor of 100,000. But Black had a difficult time convincing others of the merits of his invention. At the time feedback was largely considered undesirable by electrical engineers. Feedback could cause an amplifier to “sing” and start generating its own output, known as self-oscillation, overwhelming the input signal. (Think of the high-pitched sound that you get when placing a microphone next to a speaker.) Engineers went to significant efforts to prevent feedback-related problems.

At the time it was also believed that an amplifier with high levels of feedback would be fundamentally unstable. Opposition to Black’s amplifier was so severe that securing a US patent required “long drawn-out arguments with the patent office,” and the British patent office treated the invention the way they treated perpetual motion machines, demanding a working model. Harold Arnold, who had since become director of research at Bell Labs, “refused to accept a negative feedback amplifier, and directed Black to design conventional amplifiers instead.”

In practice, keeping the amplifier stable (avoiding self-oscillation) while also stringing together amplifiers in sequence proved to be a complex problem. These issues were resolved in part thanks to the help of two other Bell Labs researchers, Harry Nyquist and Henrik Bode. ...
Click to expand...

Next year we celebrate 100 years of negative feedback amplifiers, or at least those of us that aren't scared of negative feedback can do so.
 
solderdude said:
There are actually transducers that have 'mechanical' vibrations (and can pick those up) in the 1-22MHz range.
Used for medical echo's.
The air is too 'thick' for those to be able travel through air (hence the stuff they put on the skin) so are not audible but they do exist.

Still waiting for the first audiophile to put one on top of their speakers and claim a super-duper tweeter able to reach 20Mhz and make all kinds of claims. :)
Click to expand...

I might sell audiophile acoustic gel. $195.99 / oz
 
TubeDesigner said:
At the power amplifier company where I used to work, we rated one of our amplifiers in horsepower. Seriously - sounds like something Cerwin-Vega would do, but there you are! ;)
Click to expand...
I believe one of Bob Carver's companies' amplifiers was rated in hp. That's an easy conversion: 1 hp = 746 watts (give or take, but that's what I was taught). I don't think that there is is an SI reference standard horse, though, so there is that. I guess I am also not sure about the precise definition of a root mean square horse as opposed to a peak to peak horse. ;)
EDIT: oops, looks like my memory's faulty yet again :facepalm: Did a quick search @WRH and turned up (only) this (Soundscraftsmen):
1774837635629.png

source: https://www.worldradiohistory.com/h...0070.pdf#search="carver amplifier horsepower"
 
Last edited:
TurtlePaul said:
That is not quite true - you make too strong of a statement. The chart of feedback level vs. harmonic distortion has already been posted.

In designs with <20 dB of negative feedback, the result is a reduction in overall distortion but the mechanism is to trade second harmonic, which is the name note on octave up, for third harmonic and higher order harmonics.

This isn’t an issue for the Topping amps which are using 100 dB+ of negative feedback, but you can make amps sound worse by tuning the feedback to maximize third the fifth harmonic.
Click to expand...
Yes, it's an interesting chart, but mathematically unsound. If you approach this from a mathematical control theory perspective, you struggle to draw the same specific conclusions. I'm surprised that the author does not knows this...
 
TurtlePaul said:

Here is Bruno Putzeys making the same chart (Fig. 14) so I assume it to be true.
Click to expand...
"negative feedback around simple nonlinearities creates distortion components that weren’t there before"
As I stated, if you construct a sufficiently requisite feedback / feedforward architecture, this doesn't emerge.

"Trying to solve analytically an integrating control loop that has a dependent error in it is impossible."

...

"The story of re-entrant distortion has an unambiguous conclusion: there simply is no such thing as “too much” feedback. There is only something as not enough feedback and it happens to be exactly what so-called moderate audio designers call “modest amounts” of feedback."
 
Going back to the early 1970s, the more feedback the better was what I was taught, but of course, as Putzey's excellent paper says, it's doing it right that's the difficulty. Phase shift turning negative into positive feedback was always the issue with valves, due to AC coupling and the inevitable output transformer in power amps. SS had less of the issue, but still needed attention. Interesting (at least to me) is that a unity gain buffer made from a single opamp like the 5532 has a vast amount of feedback, even at 20kHz, yet is extremely stable even into capacitative loads.

S.
 
Axo1989 said:
Yes it’s odd reading this thread without knowing who/what the OP is referring to.
Click to expand...
DNCAgain said:
My statement was incorrect, I apologize.

In fact, his description is completely the opposite. The advantages he claimed are as follows:

1. Pure digital design (no need for digital-to-analog conversion);
2. Feedback-free design;
3. GaN material (provides greater thrust);
4. PCM to PWM framework.

However, he did not show any data except for labeling the power as 500W×2!
Click to expand...

Ok you bottom-quoted my post, but unless I missed a post somewhere I've still no idea who 'he' is, or what the material is that you refer to.
 
Maybe it's about this amp:

audioxpress.com

GaN Systems and Axign Demonstrate New 500W Class-D Reference Audio Amplifier

GaN Systems, the leading GaN power semiconductor specialist and audio technology innovator Axign, debuted a new groundbreaking GaN-based 500W Class-D audio amplifier. This reference design merges best-in-class technologies, including GaN Systems’ GaN power transistors and Axign’s Class-D...
audioxpress.com audioxpress.com

Which is certainly not feedback-free!
 
Axo1989 said:
Ok you bottom-quoted my post, but unless I missed a post somewhere I've still no idea who 'he' is, or what the material is that you refer to.
Click to expand...
Below is the opening post. The "he" is most probably the someone that has a website on the net.
DNCAgain said:
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.”
Click to expand...
 
MaxwellsEq said:
As I stated, if you construct a sufficiently requisite feedback / feedforward architecture, this doesn't emerge.
Click to expand...
“There's absolutely nothing in the theory that would justify anybody making a claim that the feedback or feedforward loop makes things sound worse. Any engineer attempting to make such a claim should back it up with proper mathematical analysis.”

You did not state what you claim to have stated. I quoted directly a false statement you made. Nowhere in your post did it cite “sufficiently requisite” feedback.

Then you claim that chart showing rising high order distortion for low levels of negative feedback was false - despite being well established in research going back to Bell Labs.

If you are going to cite control theory, then it on you to be precise in your description.
 
You must log in or register to reply here.

Similar threads

Michael Fidler
SPARTAN 5 design files and technical description - a very simple phonostage that does it all!
Replies
8
Views
3K
Michael Fidler
Michael Fidler
yo!
The Topping E8x8 Pre in practice. A few thoughts and first impressions.
Replies
19
Views
4K
1231rq32r1qw32r
1231rq32r1qw32r
amirm
  • Locked
  • Sticky
  • Featured
Audio Science Review FAQ #1
Replies
0
Views
20K
amirm
amirm
Back
Top Bottom