Amps
[Top] [All Lists]

[AMPS] G2DAF

To: <amps@contesting.com>
Subject: [AMPS] G2DAF
From: W8JI@contesting.com (Tom Rauch)
Date: Thu, 15 Jul 1999 11:20:24 -0400
This fits well with the topic of people not knowing everything, and 
making mistakes.

> >> is Maxwell's conjugate matching theorem.  If it were 100% true,
> >> amplifiers could be no more than 50% efficient.  .  
> >
> >And the logic behind that statement is???

Rich wrote:

> ?  Power transfer is maximal when the resistance of the load equals the
> internal of the generator.  At this point, efficiency is 50%.   However,
> in practice amplifiers are c.60% efficient.  

You've made the common mistake of looking at "pictures in books" 
and not reading all the text explaining the pictures.

The rules of the model you are trying to use are:

1.) The model can never be used to explain actions in the source. 
Thevenin was very clear about this. It only describes the box as a 
two terminal black box with a non-dissipative resistance and a 
PERFECT zero impedance voltage source..

2.) The source must be linear (by that they mean it has to have the 
same characteristics as power is varied) in order to be modeled 
that way.

3.) The Thevenin model can be FREELY interchanged with a 
Norton model that uses a resistance in parallel with a perfect 
current generator by changing resistance parameters to admittance 
form.

If you do the interchange you will find your 60 percent amplifier 
model made by wrongly applying Thevenin's theorem to a PA 
suddenly becomes less than 50% efficient with the very same 
parameters of operation.  
  
The maximum power transfer theorem clearly states maximum 
power is transferred to the load when the system is conjugately 
matched, and that if a system is conjugately matched at one point 
it is conjugately matched at all points in the linear portion of that 
system. None of this has anything to do with dissipative 
resistances, and only a person who never read the theorems would 
think it did or that it could be used to describe efficiency in a PA.

Breaking two rules of a model is not a good way to illustrate you 
understand the model.

> >It's a bad idea, for a multiband amplifier, to increase unwanted 
> >feedthrough. 
> 
> ?  The hat trick is to move the  grid's parallel-resonance (where 
> feedback/feedthrough is maximal) away from the anode-resonace, where
> damped wave VHF ringing takes place whenever the anode current changes. 

There are several flaws with that theory.

First, grid resonance is caused by the self-capacitance of the grid 
resonating with the grid lead inductance. That self-capacitance is in 
the order of a few pico-farads. If you add an EQUAL amount of 
capacitance in series with the grid you shift the resonant frequency 
LESS than 50% higher. That would require about 10 pF or so in an 
811A.

Adding a few hundred pF barely changes the self-neutralizing 
frequency of the tune, when that parameter is properly measured. 

Second, adding series reactance INCREASES feedthrough one 
every other frequency. Collins (and Bill Orr when he "borrowed" the 
idea from Collins) claim this is to add "negative feedback".

That's pure nonsense, as anyone who calculates the grid voltage 
produced by the capacitive divider formed by the cathode-grid 
impedance and grid-to-chassis impedance would be able to see.  

> >Especially when the tubes have too much feedback to 
> >be stable without neutralization when the grids are grounded.
> >
> ?  In the world of VHF, a grid can not be absolutely, positively brought
> to ground potential.  > -  later, Tom

True.  But they can be neutralized at and near the operating 
frequency range and proper negative feedback can be added in the 
cathode.

A good book is Circuits and Networks by Koehler. Read the 
section on matching theorems.


73, Tom W8JI
w8ji@contesting.com

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm


<Prev in Thread] Current Thread [Next in Thread>