I think Gary Schafer's analysis (below) is basically correct but I have
minor comments to add (in addition to my earlier post, also quoted below,
claiming the 400+ peak output just could not be possible, but I think I
was wrong about that). See below.

On Tue, 24 Jan 2006, Gary Schafer wrote:

On Tue, 24 Jan 2006 16:07:30 -0800, Don Bowey
wrote:

On 1/24/06 2:31 PM, in article ,
"Gary Schafer" wrote:

On Tue, 24 Jan 2006 13:40:54 -0800, Don Bowey
wrote:

On 1/24/06 12:57 PM, in article 6xwBf.11951$bF.2404@dukeread07, "Uncle
Peter" wrote:


"Straydog" wrote in message
My understanding of AM transmitter technology would estimate that a 32v3,
with ~120 DC input (two 6146s, or were they still using one 4D32?) would
have at most (class C, plate modulated) 70% X 120 = 80 watts of CW carrier
output. 60 watts of audio on that final tube (as a non-linear high level
mixer) will at best, double the _instantaneous_ (peak) input voltage,
therefore power to 240 watts (plate current will _not_ double even if the
plate voltage doubles on peak audio cycle [look at your tube curves again
of iP vs vP at constant biases]) which you could only attempt to measure
with an oscilloscope. Peak output? Could it be more than 240 x 0.7 = 168
watts? I doubt it (unless he's got something like "super-modulation" in
the rig).


Without delving into the limitations of the 32V3, according to the info
from an ARRL publication:

"..since the amplitude at the peak of the upswing is twice the unmodulated
amplitude, the power at this instant is four times the unmodulated, or 400
watts."

Average power, on the other hand, will be 1.5 times carrier. A Class C
amplifier with high level modulation should produce an instaneous PEP
of 4x carrier power.

Pete






Getting back to basics: A 120W (input) power, class C stage, will require
60W of audio (using a high-level, e.g. plate, modulator) for 100%
modulation. If we assume 85% efficiency, then the output will consist of a
Carrier of 102W and two sidebands of 25.5W each.

In my opinion, any other explanation is useless. Do remember that the
carrier amplitude does NOT vary with modulation.

Don


I don't remember the 32v3 specs but a pair of 6146B's is rated for
120 watts carrier output on AM. 6146A's are rated for 100 watts output
on AM.

My Viking had a 4D32 final and it would load to well over 100W.


Assuming the 120 watts carrier output, when modulated 100% the voltage
doubles and the current also doubles on modulation peaks. Doubling the
voltage and doubling the current works out to 4 times the power. This
is of course Peak Envelope Power of the signal which would be 480
watts.

Where does the double voltage come from at 100% modulation? I can only
account for a 50% rise in voltage.


You can not just add the audio power to the carrier power to find PEP.
You must first add the voltages together.

Good idea, if one knows the voltages......


Peak envelope power is what the FCC is concerned with for maximum
allowable power of 1500 watts.

Although when advertising an AM transmitter it is common to state the
carrier power and not try to confuse people by stating the PEP power
and not stating that is what is being speced.

73
Gary K4FMX


My point is that listing the PEP capability of an AM transmitter isn't as
useful as stating it can output about 100 watts.

Don


I agree stating PEP output of an AM transmitter does little.

I also think it doesn't mean much even for an SSB signal (which is
difficult to compare with AM coming from the same station) because the
S-meter damping makes it difficult to measure signal strength. Also, for
the power in two sidebands (only one of which is needed) and the waste in
the carrier, the usual efficiency of a linear amp is about half of that
for a (non-linear) AM final amp.

But a
properly operating transmitter should be able to give pep at 4 times
the carrier power.

I think this contradicts something you said below, and contradicts what I
said in my post, above. See more below.

Some transmitters do not have that capability
because of a poor modulator or too small finals, or power supply etc.
I, like Peter, was trying to dispel the somewhat misleading add of the
original poster.

As far as the voltage doubling with modulation, you only need to look
at the output on an oscilloscope at the composite signal and you will
easily see that it does. Set the scope to show the carrier level at
say 2 divisions on the screen. With modulation you will see the
positive peaks reach 4 divisions on the scope. The negative peaks will
reach zero on the scope.

Yes, and I have done this, myself and seen a carrier "band" on my scope,
and when speaking into the microphone (on a Johnson Ranger), seen the high
peaks go up to about double the height of the carrier and the valleys go
down to about zero (below zero would take the carrier away thus leading
to splatter).

Another way to look at this is when modulating the final the peak
audio voltage must equal the plate voltage for 100% modulation. In
order for the modulation to go to negative 100% the audio voltage must
cause the plate voltage to swing down to zero. By the same note in
order to reach 100% positive modulation the audio voltage must cause
the plate voltage to go to twice the dc voltage.

Yes, but none of this explains where the "4X pep" statement comes from. In
fact, even at the instantaneous double the plate voltage, there is no
plate current increase. The (non-linear) tube is not a (linear) resistor
where you double the voltage accross the resistor and cause the current
to double, thus a quadrupling of power. Look at the curves in your tube
manuals for any given control (triodes if no other grids are present) or
screen grid (tetrodes or pentodes) bias. Above some threshold plate voltage
the plate current is independent of plate voltage. Plate current is only
affected by grid voltages.

It may seem confusing because if you add the average output power up a
100 watt transmitter is only 150 watts. 100 watts carrier and 25 watts
in each side band.

I think this is right.

However if you add the voltage of the carrier plus
the voltage of each audio side band and then calculate the power you
will see that it is 4 times the carrier power.

I don't think this is quite right and, after thinking about all of this,
part of the reason I already gave above is also not quite right.
Here is another way I think we can look at this question (see below):

100 watts into 50 ohms = 70.7 volts
25 watts into 50 ohms = 35.35 volts
25 watts into 50 ohms = 35.35 volts
Total voltage = 141.4 volts (which is 2 x carrier
voltage)

I think this is a bit of a mistake and it would be better to calculate
peak power in the following manner:

First, there is no peak output power from the carrier, the carrier is
always there and at the same strength no matter if there is modulation or
not.

So, power contribution FROM THE CARRIER (whether modulated or not) is
still only your 100 watts, period. The carrier contributes NO extra power
to the peak power we're all interested in.

Now, lets look at the power in the sidebands. I'll accept that 25 watts as
converting to 35 volts, but that 35 volts is not added to carrier
power because it is in the 2-3 kc spectrum above or below the carrier.
So, use your formula below and get (35 X 35)/50 = 1225/50 = 24 watts.
And, for the second (other) sideband, there will be another 24 watts.
Total: 48 watts of audio translated to RF in addition to the 100 watt
(constant) carrier. Thus peak power is 148 watts.

Where is the conflict between my analysis and yours?

Its in the way we think about modulator power output (usually stated as
audio power must be about half of final DC input, thus a 120 w DC input
class-C final needs about 60 watts of audio). So, when you look up, for
example the specs on a pair of 6146s in modulator service (go look in the
back of your ARRL handbooks, any of them) and see them talk about 110-130
watts for the pair in either AB1 or AB2! They don't tell you that is _peak_
audio power! For the same tube in class C, they are showing 50-70 watts
(continuous RF) out for one tube. Not much difference in power specs per
tube, but you won't get anywhere near those 110-130 watts of audio in
continuous (i.e. average) power because the heat dissipation will melt
the plates since class AB is much less efficient than class C.

So, your example of 25 watts per sideband is more like an _average_ power
specification and what we should be looking at is what is the peak
audio power (or voltage) coming out of the modulator. That peak audio
voltage out of the modulator has to be equal to the plate voltage and in
the same direction to double the final amp plate voltage, and valley
bottom audio equal to the final amp plate voltage but in the opposite
direction to reduce plate voltage to zero or near zero. Final B+ voltage
plus the audio peak thus shows up on the scope, transiently, as RF output
voltage at double the height of carrier alone, and final B+ minus the
negative audio peak, at the negative peak, causes the height of the scope
trace to go, transiently, to zero. So, when THEY talk about 60 watts of
audio power to modulate a 120 w DC input final amplifier, they are talking
about more like 60 watts average power which really means something like
120 watts of peak power (and is in all of the tube manuals where specs
for all of the amplifier classes are shown next to each other! [this is
not the case for receiving tube manuals which talk about average signal
output for tubes like 6L6s in classes A and maybe AB]). This gets us into
the audiophiles' endless arguing about what audio power means and under
what conditions and specifications (eg. the distortions) need to be made
for considering peak audio power, especially in music audio (with eg. drum
beat transients) rather than voice audio (more or less steady).

My general feeling is that in the AM transmitter situation (and the SSB
situation, too) that talking about peak power has only theoretical value
and almost no practical value and is confusing. You can only measure it on
a scope and S-meter readings will be subject to the damping factor in the
mechanical needle, electronic fudging by so-called "peak reading" meters
including the bar graph things, and any asymmetry in the voice waveform,
and distortions and non-linear characteristics in the rest of the
electronics. I have heard guys on the air using the same amplifier in a
linear class mode who switch from AM mode to SSB mode and my S meters (on
lots of receivers) show the same peak value on SSB as the steady value on
AM, plus or minus maybe one or two db, at most.

Guys would be best off talking about final amp measured DC input watts
(continous) and/or final amp measured carrier (continuous) RF output watts
and not say too much about their modulator power unless they have a scope
on the instantaneous modulator output voltage and current and can actually
make a real, valid, representative measurement of both peak and average
power and be able to say "Oh, my actual, real, measured-on-a-scope
instantaneous peak whatever is X plus or minus Z accuracy."

Now, how about peak power input to your 100 watt incandescent light bulb
at home? Remember your house AC line voltage (117 VAC) is measured and
speced as RMS (root mean square), so peak voltage is 1.41 X 117, peak
current is 1.41 x 1 amp, and peak power is thus 1.41 X 1.41 X 100 watts?
About 200 watts? Is that meaningful? No, because it isn't. Think DC and
why RMS specs are always used for AC circuits and VOM voltmeter scales.

Thanks for your attention, sorry to be long-winded.

Art, W4PON

P= E squared / R

141.4 x 141.4 = 19994
19994 / 40 = 400 watts

73
Gary K4FMX