[AMPS] Conjugate Matching In Class B and C Amplifiers

[Tom Rauch]

> Conjugate matching is not required for unconditional stability.    That's
> incorrect.  Many low noise amplifiers are purposely mismatched because
> simultaneous conjugate match generally does not equal minimum noise
> figure.

Conjugate matching deals only with the condition for maximum 
available power transfer.
 
> Additionally, adjusting input match does not necessarily always affect the
> output match and vice versa.  I am trying to think whether or not your
> statement is correct.  Too many cobwebs upstairs right now.

Input matching only affects output matching when the amplifier has 
feedback of some type.
 
> Generally speaking of this subject, I had always thought that the
> efficiency of an amplifier and whether or not you simulataneously match it
> were independent.  It wasn't until coming onto this reflector and reading
> stuff from the likes of Dick Erhorn that some people feel that a conjugate
> match guarantees 50% efficiency or less.

Not in any of my books. That falsehood assumes the source is 
entirely limited by a dissipative resistance.

If I go out in my field and put a tuner between an antenna of 450 -
j300 ohms and a 50 ohm feedline that delivers maximum power into 
a 50 ohm j0 load, and adjust the tuner for maximum power into the 
antenna it will be nearly perfectly conjugately matched. The only 
error will be caused by system losses.

It's amazing how we all lived happily with this concept for many 
years and now suddenly we are to believe the system becomes 
50% efficient when matched.

Norton's Theorem, where a perfect current generator is placed 
across a source resistance,  plainly states it is 100% 
interchangeable with Thevenin's Theorem (where a ideal voltage 
source is connected in series with a source resistance). Think 
about that a while. 

The men who set the rules of the Theorems said the models are 
100% interchangeable. The Theorems also state such theorems 
can not be used to determine anything more than maximum energy 
transfer, and that they can't tell us anything about the non-linear 
part of a system or anything about source efficiency.

It is silly to use a model outside the limits defined for the model as 
an argument that the model is no good.
 
> So some people here showed me the view on efficiency being 50% or less; I
> stated it in an earlier post; Tom Rauch says that is incorrect.  I haven't
> had the time to go back to my textbooks and look at what they 
>say.

Get a copy of Circuits and Networks by Koehler. Read chapter 2, 
especially 2-5 (Maximum Power Transfer Theorems).

What has happen here is people have taken a model that clearly 
has limits and tried to use it to describe things that the model says 
it can not be used to describe. All these models do is describe the 
condition for optimum power transfer in four terminal networks.

Now we have a bunch of people misusing the models to predict 
things (source efficiency) that the models clearly state they can 
not do.

One proof the models are misused is the fact that Norton's 
Theorem states the Norton model ( a perfect current source with 
infinite impedance across a source resistance) is 100% 
interchangeable with a Thevenin model (a perfect voltage source in 
series with a source resistance).

Yet the fellows who say the models prove efficiency say the source 
would have 100% efficiency only if source impedance was zero by 
showing a resistor in series with a generator. Since the Norton 
model is interchangeable, the 100% efficiency condition would 
become ZERO of we converted to a Norton equivalent! That's what 
happens when people start thinking models are real-world 
transmitters, and start using the models to describe something 
that the rules of the model say can not be described.

A conjugate match means we have adjusted impedance in the 
system so we are extracting optimum power from a source. 
Nothing more, nothing less.

This is simply a bunch of people arguing about nothing, and 
misusing the models to prove they are "right". We all know we have 
optimum power transfer when the system is matched optimally to 
the impedances, and that lossless linear networks are bilaterial. 
That means if the system is matched in one direction, it is 
matched in the other. That's a conjugate match.

The system certainly doesn't HAVE to be conjugately matched, 
and if it is the conjugate match has NOTHING to do with efficiency 
being 50% or any other value. 

It's one of the silliest arguments ever, and will go on forever 
because the models are being misused. It appears some people 
actually think the source IS a perfect generator in series or in  
parallel with a real RESISTOR.
73, Tom W8JI
W8JI@contesting.com 

--
FAQ on WWW:               http://www.contesting.com/FAQ/amps
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com