On Jun 12, 11:56*am, K7ITM wrote:
On Jun 12, 8:17*am, Richard Fry wrote:

On Jun 12, 9:44*am, Wimpie wrote:

A real class C amplifier with very small conduction angle has
efficiency 50% when optimally tuned.

Would that not be evidence that the amplifier source impedance
necessarily is lower than the load impedance?

No, not at all. *It is important to realize that a real source is NOT
necessarily like either the Thevenin or the Norton equivalent. *The
Thevenin and Norton equivalents, when discussed in any respectable
text, will be noted to behave exactly like any other real linear
source with respect to the load, but NOT necessarily with respect to
the source itself, and in particular with respect to source efficiency
and dissipation.

For example: *I can, at least theoretically, build a switching power
supply (a source) which measures its output voltage and current and
adjusts the voltage down by 50 volts for each amp that's drawn from
it. *Because it is always operating as a switching supply with very
high efficiency, even when it's loaded so the output voltage is half
the zero-current output voltage, it's not behaving internally like a
Thevenin equivalent, even though it has a 50 ohm output resistance.





When it operates at the
transition of current/voltage saturation is can show 50 Ohms output
impedance for very small change in load impedance. But as soon as the
load mismatch is above about VSWR = 1.05…1.1, output VSWR/impedance of
the amplifier changes rapidly.

For some additional input -- I have taken part in factory tests of
high-power Class C single-tube/tuned cavity FM broadcast transmitters
driving 50-ohm test loads measured to have 1.03VSWR, *showing a DC
input to r-f output efficiency of the PA to be in excess of 80% --
including the loss in the harmonic filter. *Load power was measured
using calorimetric methods. *In fact, 80% PA efficiency is the
published spec for this transmitter line as long as load VSWR relative
to 50 ohms is 1.7:1 or less (any phase angle).

Those results don't appear to be fitting very well with the idea that
the tx source impedance can be ~50 ohms for loads with low VSWR.

RF

As Wim says, the conjugate-matched output impedance is guaranteed only
over a very narrow range of conditions. *Specifically, it's for the
condition where you can't adjust the load impedance in any direction
without lowering the power delivered to the load. *For a power output
stage with a tank that can tune both resistance and reactance (such as
a pi network), that should be equivalent to adjusting the pi network
to maximize the power delivered to the load with a particular fixed
set of bias/drive conditions for that output stage.

But why would I necessarily want to operate the output stage that
way? *What if I'm providing enough grid drive and plate voltage to my
6146 that it's capable of outputting 150 watts (class C), but if I
were to tune up the output matching network to get that much output
power, the tube's life would be significantly shortened because the
plate dissipation would be too high? *What if I'm operating it class
AB, with low enough drive so that the plate voltage and current change
only by small percentages?

Or to me, more to the point: *why would I care what the source
impedance of my transmitter is? *With the power at the wall outlet in
my house, I expect a nearly constant voltage, and have no intention of
loading it with a "conjugate match." *I don't know WHAT its source
resistance is, exactly, but I know it's much lower than the loads I
put on it. *With an electric motor, I have no intention of putting a
load in excess of the motor's rating on it, for fear of burning up the
motor, even though I know that for short periods the motor can deliver
quite a bit more output power than its nameplate rating. *With a
switching power supply, same thing: *I don't load it to the point of
significant drop in output voltage. *All I want from ANY of these,
including the RF power amplifier, is the ability to deliver the
expected power to the rated load.

In the instrumentation systems I deal with professionally, there are
times I care very much about source impedance. *With respect to ham
transmitters and power amplifiers, though, I just can't get excited
about caring that much what the source impedance is. *The point of the
amplifier or transmitter is to deliver power to my antenna system, and
the source impedance it represents is irrelevant to that task.

Cheers,
Tom

Hello Richard Fry,

I quote a sentence from your previous post:

"Would that not be evidence that the amplifier source impedance
necessarily is lower than the load impedance?"

What you have said above is the key to the concern over the output
resistance of a Clsss C amplifier being non-dissipative. What seems to
be universally misunderstood is that there are really two separate
resistances in the operation of these amps; one, the cathode-to-plate
resistance, which is the dissipative resistance Rpd that accounts for
all the heat, due to the electrons striking the plate; and two, the
real output resistance that appears at the output of the pi-network,
the resistance comprised of the voltage-current ratio E/R. I am
speaking only of tube amps using an adjustable pi-network in the
output. Due to the energy storage in the pi-network the output of the
amp is inherently isolated from the upstream non-linear portion, thus
the output is linear. One of the myths concerning the RF amp is that
it cannot have efficiency greater than 50% because half of the power
is dissipated in the source resistance. This would be correct if the
output source resistance was dissipative, like in a classical
generator. But in the real RF xmtr this is not true.

Returning now to your quote above, your reference to the source
impedance (resistance Rpd) is actually the plate-to-cathode
resistance, the dissipative resistance, which is less than the output
resistance R = E/I appearing at the terminals of the pi-network. I
must add that when the pi-network has been adjusted to deliver all the
available power at any particular grid-drive level, the output
resistance of the network equals the load resistance, but not
necessarily 50 ohms, i.e., whatever the load resistance might be.

If you have doubts concerning the conditions I stated above, I invite
you to review Chapter 19 in Reflections 3, which is available in two
parts on my web page at www.w2du.com, as the first part appears in
Reflections 2 and the second part appears in 'Preview of Chapters in
Reflections 3'. In that chapter I use an example taken directly from
Terman's 'Radio Engineers Handbook'. You will note that the
dissipative resistance Rpd (3315.6 ohms) is less than the output
resistance (6405 ohms) obtained from the E/I relationship appearing at
the output

This example verifies the explanation in the text. In addition, I also
invite you to review the portion of the chapter appearing in the
'Preview Chapters in Reflections 3', in which I report additional
measurements that offer additional proof that the output source
resistance is non-dissipative. The numbers prove it to be true.

Walt, W2DU