Hi,
And he is simply wrong. And makes no attempt to define
optimum Class B (Selfs definition) Class aB in my terms.
Its a known fact he is confusing.
rgds, sreten.
Hi Tom
Actually what confused me was the fact your graphs are for a single distortion harmonic each. Usually a graph of this shape plots total THD vs output level, and since THD is a ratio, it will rise as output decreases.
Otherwise, low-f THD should be unmeasurable (or close to it) even for an amp with moderate open-loop gain.
Sreten - Leech defines the transconductance of the output stage as a whole, which depends on the load impedance. Within this definition there is no gm-doubling and it is a valid way to view things. With a resistive load, the transconductance is then constant and only changes if you change the load resistance.
The incremental gm and instantaneous gm of each output device at a given load current can be viewed as"internal-gm" or "raw-gm", with the whole output stage and load showing the "net-gm". Again, just many ways to assess the same situation, and different aspects to explore when trying to figure out causes for distortion and their fixes.
Leech is very open-minded if you read all of his discussions about his amp evolution. Self is much more narrower in his views.
Have fun
Actually what confused me was the fact your graphs are for a single distortion harmonic each. Usually a graph of this shape plots total THD vs output level, and since THD is a ratio, it will rise as output decreases.
Otherwise, low-f THD should be unmeasurable (or close to it) even for an amp with moderate open-loop gain.
Sreten - Leech defines the transconductance of the output stage as a whole, which depends on the load impedance. Within this definition there is no gm-doubling and it is a valid way to view things. With a resistive load, the transconductance is then constant and only changes if you change the load resistance.
The incremental gm and instantaneous gm of each output device at a given load current can be viewed as"internal-gm" or "raw-gm", with the whole output stage and load showing the "net-gm". Again, just many ways to assess the same situation, and different aspects to explore when trying to figure out causes for distortion and their fixes.
Leech is very open-minded if you read all of his discussions about his amp evolution. Self is much more narrower in his views.
Have fun
Last edited:
Trying to maintain idle current strictly null is probably impossible.
Either, there will be a small quiescent current, that's the case for most op-amps and I've always seen them specified as having a class B output stage.
Or bias is insufficient, and there is a small dead zone around 0 V where the signal is not reproduced in open loop. It can be called a class C output stage (as in the Quad 405 before error correction being applied) or each branch can be considered in class C.
I found Self's classification clarified a previously incoherent and confused topic.
Hi,
I don't think Self clarified much at all. Anyone familiar
with typical class C radio circuits would completely
disagree with Self's use of the term. Class C only
conducts on one half of the cycle, and that
meaning predates Self's musings by decades.
Class B conducts on both half cycles.
Defining class B as including optimum bias is just very wrong.
How do you know its optimum ? and not C or AB in his terms ?
The Quad 405 output is pure class B. In my book most
amplifiers are class aB, where "a" represents the bias.
Note that aB implies some bias, but not the optimum.
Class AB is biased beyond aB to a clear class A region.
That region is "Gm doubled" by definition.
rgds, sreten.
I don't think Self clarified much at all. Anyone familiar
with typical class C radio circuits would completely
disagree with Self's use of the term. Class C only
conducts on one half of the cycle, and that
meaning predates Self's musings by decades.
Class B conducts on both half cycles.
Defining class B as including optimum bias is just very wrong.
How do you know its optimum ? and not C or AB in his terms ?
The Quad 405 output is pure class B. In my book most
amplifiers are class aB, where "a" represents the bias.
Note that aB implies some bias, but not the optimum.
Class AB is biased beyond aB to a clear class A region.
That region is "Gm doubled" by definition.
rgds, sreten.
Last edited:
Hi Guys
Referring to opamp output stages as "class-B" is at best just being colloquial and at worst is sloppy. They really want to say that it runs cool, which is a good thing for a chip.
Electronics uses precise terms so that everyone will understand exactly what is meant. Most engineers and hobbyists would call an amp with a small amount of idle current class-AB because in fact it has a class-A region of operation and a class-B region - with a caveat. because the latter region does not fit the 180-degree conduction, one could justifiably call this a class-C region...
The Quad amp had a mixed output stage using a class-A amp and class-C section. It is not a single operating class amplifier.
I've never seen a pure class-B solid-state amp, although one could probably set up the Crown DC-300 in such a way. There are class-B tube circuits that function at zero idle current yet exhibit no more THD than any other conventional tube output stage.
Oliver was always referreing to output stages where some idle current is present.
Have fun
Referring to opamp output stages as "class-B" is at best just being colloquial and at worst is sloppy. They really want to say that it runs cool, which is a good thing for a chip.
Electronics uses precise terms so that everyone will understand exactly what is meant. Most engineers and hobbyists would call an amp with a small amount of idle current class-AB because in fact it has a class-A region of operation and a class-B region - with a caveat. because the latter region does not fit the 180-degree conduction, one could justifiably call this a class-C region...
The Quad amp had a mixed output stage using a class-A amp and class-C section. It is not a single operating class amplifier.
I've never seen a pure class-B solid-state amp, although one could probably set up the Crown DC-300 in such a way. There are class-B tube circuits that function at zero idle current yet exhibit no more THD than any other conventional tube output stage.
Oliver was always referreing to output stages where some idle current is present.
Have fun
- Status
- Not open for further replies.