[AMRadio] AM Linear Efficiency (was Ongoing Quest)

[D. Chester]

The efficiency of a linear amplifier is directly proportional to the 
amplitude of the signal it is handling, between the two extremes of (1) no 
signal and (2) the saturation point, where any further increase in driving 
power results in no further increase in output power.

Under "no-signal" conditions, the power input to the linear amplifier 
consists only of  plate voltage X static plate current.  Power output is 
zero.  Therefore the efficiency is zero.

At saturation, the efficiency of a class-B linear amplifier is typically 
somewhere around 60%, and may well be considerably less once additional 
losses including those of the output tank circuit, antenna tuner and feed 
line are factored in.

Between those two extremes, the efficiency is directly proportional to the 
signal level.  It doesn't matter whether the signal is AM, SSB, CW, FM or 
any other mode.  With AM, the steady carrier cannot drive the amplifier to 
the full saturated output point, since sufficient headroom must be left over 
after the carrier to accommodate the positive modulation peaks. If the 
amplifier is driven to full saturation right at the 100% positive modulation 
peaks, the unmodulated carrier level must be set to a point exactly midway 
between zero and saturation. The RF carrier voltage is therefore half the 
voltage at saturation, and the efficiency of the amplifier stage is likewise 
half of what it is at saturation.  Using the 60% figure for maximum 
efficiency, the carrier efficiency with no modulation present is half that, 
or 30%.  Exceeding 30% resting carrier efficiency results in flat-topping at 
positive peaks, which generates exactly the same kind of distortion and 
splatter as overmodulation in the negative direction.

Interestingly, the plates of the tubes will dissipate less power and run 
cooler with modulation present, than with no modulation.  If the stage is 
really a LINEAR amplifier, the DC input to the final must remain perfectly 
steady with or without modulation.  The plate current should NOT shift 
upwards or downwards when you talk.  Since the DC input is steady, but with 
modulation additional power appears in the output as sideband power, that 
additional power has to come from somewhere.  The only possibility is for 
the efficiency of the linear output stage to increase, so that some of the 
DC power input is converted to sideband power rather than heating up the 
plate of the tube.  If the tube is one of the types whose plate glows red at 
normal plate dissipation levels, the plate glow will DECREASE with 
modulation, even though the output power increases.

The greatest advantage with class-C plate modulation is that the final stage 
runs at saturation at all times.  With AM linear, the drive to the amplifier 
and the plate loading are very critical, in order to maintain the point 
where the unmodulated  carrier is exactly at the midpoint between zero and 
maximum output. And the modulation level must be adjusted to just barely 
modulate 100% but not overmodulate.  OTOH, with plate modulation, it is only 
a matter of making sure the grid has enough drive to maintain class-C 
operation (saturation), and that the modulation level adjusted to avoid 
overmodulation at the negative peaks or flat-topping at the positive peaks. 
Usually, all you have to do with plate modulation is make sure the plate is 
dipped, the grid drive is peaked and running somewhere close to the normal 
level, and that the plate loading is set for normal operating plate current, 
and (using an oscilloscope or modulation monitor) that the transmitter is 
not being overmodulated or driven beyond its modulation capability.

True, the resting carrier efficiency is much greater in a plate modulated 
stage than with a linear amplifier stage, but with plate modulation you have 
the additional power consumed by the modulator tubes (plate, filament and 
possibly screen grids) plus all the power consumed by the audio driver 
stage. So the overall efficiency between the power drawn from the a.c. mains 
and the power input to the antenna turns out to be close to the same for 
both methods, or at most, SLIGHTLY better with plate modulation.

Grid modulation, whether control grid or screen grid, is a form of 
efficiency modulation and works very similarly to linear amplification, with 
the same critical tune-up precautions. In fact, with control grid 
modulation, you can think of the rf grid drive as being steady while the DC 
grid bias is varied, whereas with linear amplification, the DC grid bias is 
held steady while the rf grid drive is varied - two very similar situations 
with essentially the same end result.

Another point to ponder is that, since the efficiency characteristic of a 
linear is exactly the same  whether it is amplifying AM or SSB, linear 
amplification of SSB suffers from the same low average efficiency as does 
AM.  With SSB, the efficiency is at its maximum only at voice peaks, just as 
with AM. With normal human speech, the average amplitude is around 35-40% of 
the peak level.  Ever notice the behaviour of a mechanical analogue VU meter 
on broadcast equipment? While it may hit the red zone on occasional peaks, 
most of the time it rides somewhere around -7 on the scale.  That represents 
the average amplitude level of an uncompressed/unclipped speech waveform. 
That means  the average efficiency of the SSB linear is running at about 40% 
of the way between zero and 60%, or about 24% efficiency  total.  What makes 
the noticeable difference is in the duty cycle.  With AM, the carrier runs 
at 100% duty cycle, so the average power dissipated in the plates of the 
tubes is considerable, while SSB runs at a very low duty cycle, so even 
though the amplifier still averages very low efficiency, the average power 
dissipated in the tube plates is much less than with AM.  That explains why 
the early "sweep tube" SSB linears would go into Chernobyl meltdown mode 
with a sustained whistle into the mic.

If you are seriously looking for "efficiency" in an AM transmitter, read up 
on "Class E" technology.


Don k4kyv


_______________________________________________________________

This message was typed using the DVORAK keyboard layout.

http://www.mwbrooks.com/dvorak/
http://gigliwood.com/abcd/