Re: How to explain how negative feedback lowers noise?

Tom Lee

Hi Ted,

I recommend continuing to read papers beyond those from 1923. It's hard to do better than to read Black's own 1934 paper (January issue of bstj, also available at, as well as Far from being a "canard", you will see from Black's own pen what the motivations for negative feedback were. If you still wish to refute him, you'll have to reanimate his corpse. The 1923 paper you cite enumerates many problems facing transcontinental telephony at that early time, but if you examine the list carefully, you'll see that every one of those problems is solvable, if clumsily, by open-loop means, with the conspicuous exception of distortion.

It's also important to recognize that the 1923 paper predates the advent of carrier-based telephony by several years (the author is still talking about loading coils!). Once multiple, carrier-modulated telephone signals were to share a circuit, the problem of distortion became the showstopper. Negative feedback solved that problem (and a host of others conveniently as well). Without negative feedback, AT&T could not have made transcontinental telephony practical.

-- Cheers,

Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070

On 3/27/2021 09:39, Ted Rook wrote:
While as a self-taught amateur physicist I hesitate to challenge a university professor but if
my understanding is correct (and it is often not so) the argument that the purpose of the
introduction of NFB was to reduce distortion should be viewed as a canard that belongs in
the realms of imagination where high-end audio fanatics can often be found.

Some insight into the background of problems with long distance (but not transcontinental)
telephony before the development of the negative feedback amplifier is provided in Bell
System Technical Journal 1923, January, author Clark, title "Telephone Transmission over
long cable circuits" which is to be found among other places as a free download at the
Internet Archive

The information presented makes it clear that the idea of reduction of distortion as the driver
of development of the negative feedback amplifier fails to recognize difficulties of a more
fundamental nature that inhibited development of transcontinental telephony and demanded
a solution, which was eventually delivered by the new invention. In addition to solution of the
fundamental problem of gain instability under hostile conditions, among the other benefits
delivered by the negative feedback amplifier were improvements to frequency response,
input and output impedances and also harmonic distortion.

The article is a fascinating insight into the very early days of long distance telephony and the
content is mostly non-mathematical.

Ted Rook

On 26 Mar 2021 at 2:03, Tom Lee wrote:

The reason it didn't happen for electronics until 1927 is that the preoccupation until that point was getting more gain per tube. Positive feedback (Armstrong's regenerative amplifier) was the magic elixir that had enabled the age of electronics to begin, around WWI. Early tubes struggled to achieve voltage gains of five. The first generation of EEs was thus trained to think of getting enough gain as the main problem.

It wasn't until AT&T ran into troubles with transcontinental telephony that someone had to invent a fix for a different problem: distortion. A cascade of hundreds of repeaters demanded individual amplifiers of unprecedented linearity. This necessity was the mother of negative feedback.

Black's invention required quite a change in thinking. Now he was recommending "throwing away" precious gain in exchange for reduced distortion. This went against the training of a generation of EEs. Given that, I'm not surprised negative feedback took that long to get formalized.

Although you are absolutely right that a basic intuitive notion of negative feedback was appreciated long before electronics came along, Black was the first in history to understand explicitly that excess gain could be used as currency to pay for reductions in distortion. It is as subtle a notion as it is powerful.

The very first mathematical treatment of negative feedback was by Maxwell himself, but his analysis was limited to understanding why speed governors for steam engines could go unstable. Until technology ran into the need for exquisite precision in control, there was no need for Black to come along.


Sent from an iThing; please excuse the terseness and typos

On Mar 25, 2021, at 23:14, "Ed Breya via" <> wrote
I don't know about this history, but am surprised that this didn't happen until 1929, well into the electronics era. Surely the concept of negative feedback in control systems has existed in nature, throughout human history, and in industry - at least since the steam era - look at the fly-ball governor, for example. Maybe in electronics, it wasn't so obvious, although it already existed in some forms, say for instance, with degenerative feedback in a vacuum tube, stabilizing the bias with the cathode resistor. People were driving horses, trains, planes, and automobiles successfully before 1929, using that PID controller in the skull.


Join to automatically receive all group messages.