How much does a RF amplifier use of its input power?

Question

I've been casually aware that signal gain can be added in "stages" but never really considered how that might actually work in practice. And it seems that some RF amplifiers really prefer to only "sample" the input signal rather than use any of its incoming power, e.g. some QRP amps I've seen even attenuate the input signal before using it to drive their output signal.

But for example I happened to recently see a classified listing for a Mirage B-5016-G amplifier which is sold as taking "50 watts in, 160 watts out". Does an amp like that simply allow high powered input signals, but still just dissipate/attenuate them down to essentially nothing before using only the underlying signal to gate the output current?

Or are there transistors where the input power must be relatively high in order to control an even higher power output?

Answer 1

Modern LDMOS FETs have a very high native gain — for example the MRFE6VP61K25 in its "standard test circuit" has a gain of 27dB at HF frequencies, and >20dB up to 350MHz. 27dB equals 1500W out for 3W in.

When used in an amp, they usually have attenuators on their inputs, for two reasons:

1. To ease SWR-matching of the input. Any mismatch on the amplifier's input will be reduced by the attenuator value. To use completely made up values: if the amplifier section would have an input impedance ranging from 30-j20 to 70+j40 over some frequency range, the same amplifier behind an ideal 10dB pad would have an impedance ranging from 48-j2 to 52+j4 — what was once <2:1 is now <1.1:1.

2. To comply with an FCC Part 97 rule that says that you can't sell an amplifier with more than 15dB of total gain. The same amplifier sold overseas might have a substantially higher gain, while the version sold in the US turns more of the input power into heat.

Answer 2

An attenuator pad is often used to improve the matching so that the exciter does not see a high SWR. That's already mentioned by KC2G.

Now to the other point of the question. The most common power amp configuration is common-emitter/source/cathode. In that configuration, the input power is used to drive the transistor/tube, and not much of that will pass through to the output. In other words, the system's overall power efficiency will be poor if the final stage operates at a low gain.

When designing a power amp that takes 100W and outputs 500W or 5W and outputs 20W, that final amp needs only 6 or 7dB gain, which deteriorates the overall power efficiency. Now, to deliver that level of gain with good efficiency with a bipolar transistor, a common collector (emitter follower) configuration is advantageous. This way, the power given to the base will largely pass through to the emitter, where the output power is taken. Of course, the emitter impedance will be very low; this is followed by a transformer to step up to 50 ohms. Care must be taken because emitter followers easily oscillate at a very high frequency if a parasitic capacitor between the base and the emitter creates feedback and the base impedance is not well stabilized. The supply voltage must be adjusted for the required output voltage swing (i.e., the input power and base driving impedance to maximize efficiency.