New book on amplifier design.

[syn08]

It is the high loop transmission at low frequencies that mitigates things like PSRR and DC offset. This is why, for example, one shouldn't need a DC servo with a competently-designed "voltage feedback" power amplifier.

Moreover, high forward-path gain at high audio frequencies means that there is far greater scope for use of minor loop negative feedback (local feedback) to improve forward-path linearity in a "voltage feedback" amplifier than exists with a classic "CFA" without adversely reducing the forward-path gain available to convert to the major negative feedback loop transmission necessary to improve closed-loop linearity.

This means that a competently-designed "VFA" will have a far more linear forward path than a classic "CFA" for the same major loop transmission at a given audio frequency. The aphorism "make the amplifier’s forward path as linear as possible before applying major loop negative feedback" springs to mind.

Mike,

Sorry, you are running in circles chasing your tail.

Enough said.

 

[Worldie]

And BTW, COD is by no means a property of CFAs as some like to believe (pun intended). A long tail pair VFA can very well be designed with COD, see the Stochino amp with its input stage capacitor cleverly used to dump extra current when required. Other VFA plus COD configuration exist and are liberally used in linear IC designs.

True.

 

[syn08]

Sometimes you use a CFA design because it solves some other issue that needs to be dealt with.

I don’t necessary disagree, but inquiring minds want to know: what would these issues be, in audio applications? I can tell why a CFA is ideal as a DSL line driver, but for audio, I have no clue.

 

[wayne]

Imagine a circuit that is VFA inverting input. The resistors of 10K and 100K would give a low input impedance.
By going CFA fed to the sources of the fets you can get high input impedance. Some volume control circuits would like this.
A minor difference between different lines of preamps that was never mentioned that I know of.

IC design is complicated stuff compared to most audio circuits in the high end industry. I am Happy to keep it to small amounts of transistors that match my brain size.

 

[Nelson Pass]

Wayne modestly fails to mention that his brain size is larger than mine.

I make up for it by having a narrower focus.

 

[Terry Demol]

I don’t necessary disagree, but inquiring minds want to know: what would these issues be, in audio applications? I can tell why a CFA is ideal as a DSL line driver, but for audio, I have no clue.

Just about all transformerless mic pres, discrete or mono use some form of CFB architecture.
Very low noise and distortion, even at high gains which is very freq independent.
Ideal application.

TCD