CB Amplifiers Converted to Ham

Peak envelope power is seen only on a peak reading meter. Average power is seen on a non-peak reading meter, like a standard Bird 43.

Pure 100% linear undistorted sine wave amplitude modulation has the following power relationships:

Carrier power is 25% of peak envelope power, or if you prefer peak envelope power is 4 times carrier power. A 1000-watt carrier is 4000 watts PEP with undistorted 100% sine wave modulation.

Average modulated power output is 150% of carrier power. This is because average audio modulation power is 50% of carrier power. A 1000-watt carrier is 1500 watts average power when 100% modulated with an undistorted sine wave.

RMS RF antenna current is 1.225 times unmodulated RF carrier current. A 1000-watt transmitter into a 50-ohm load is about 4.5 amperes unmodulated carrier and 5.5 amperes load current when 100% modulated.

Peak RF output voltage when modulated is 2 times unmodulated carrier voltage. A 1000-watt carrier into 50 ohms is about 326 volts peak voltage and 652 volts peak when 100% modulated.

Thinks that arc, like capacitors, must be rated at peak voltage. Things that heat are generally rated using RMS current or average power.

Amplifier Class

I'm not sure why CB amplifiers are called "AB1". Any tube amplifier with grid current is not "AB1", it is either class AB2 or class C. Most CB amps are actually class C, and there is a reason for that.

Solid-state amplifiers cannot be class  "AB1" or class "AB2". A bipolar transistor amp cannot be a sub 1 or sub 2 class because they do not have grids to have or not have grid current. Solid-state amplifiers have bases or gates, and a bipolar transistor always has base current.

Most CB amps are intentionally self-biased in at least one stage with some form of unregulated bias or supply voltage. This causes them to automatically adjust gain to compensate for differing drive powers. The primary objective is to not be linear. A linear amplifier will not "swing much" with modulation, while a grossly non-linear amp will "swing" substantially with AM voice peaks.

Generally the way CB amplifiers obtain swing is through class C biasing. While they often rest in conduction with no drive, giving the appearance of being class AB or B, they develop self-bias through a grid leak (or base leak) that pulls them into class C when enough drive power is applied. A capacitor across the leak resistance averages the bias voltage. This establishes gain at the point where output power is folded back far below the point where it would be with stiff bias. It is a form of slow time constant gain compression that sets the carrier level far below the modulation peaks. 

When an audio peak comes along they conduct heavily, causing a power meter to dance upward on voice peaks.

Another trick was an unregulated screen. They would get the unregulated screen wiggling around to accent any power change, and this would cause carrier shift.

It actually would be pretty tough to sell a real linear amp to a CB op, because it wouldn't have that great audio caused by RF compression or clipping. Most amps were intentionally or accidentally built to be RF speech processors.

Contrast that to a real linear SSB amp. In almost any real linear amp bias is very stiff. It is regulated. The screen voltage is stiff also, up to dissipation limit, in a good clean amp. 

Class AB does not mean the amplifier is linear. "Linear" in the context of amplifiers indicates a linear transfer function, or a transfer function that has the proper transfer function curve shape to minimize odd-order intermodulation products. A linear transfer function generally means the output power level tracks the input power level in a way that does not cause excessive 3rd, 5th, 7th and other odd-order mixes. Only odd-order mixing or distortion produces annoying splatter.

Often the criteria is how the "audio sounds". The on-frequency sound tells us nothing about bandwidth and nothing about harmonics.