[Elecraft] ALC overshoot

[Tom W8JI]

> leads to severe distortion or splatter. I again hold the 
> opinion that the
> real bad signals heard are due to to PAs being operated 
> with no ALC feedback
> to the exciter feeding it and operators being clueless to 
> this.

Most amplifiers have very crummy ALC detection. All they do 
is same RF level at the cathode of the tube, and that 
doesn't indicate excessive drive or nonlinearity at all! In 
most cases however people are just kidding themselves when 
they connect ALC from an external amplifier. All the ALC 
from the amp does is sample the same thing the radio's 
directional coupler does, so the external ALC is redundant 
to what the radio already has... except not as accurate or 
stable from band to band.

In some later tube amps I used an ALC detector that samples 
grid current (and in a few cases grid current and plate 
current). This does work to limit excessive drive, but in 
most radios the ALC has so much gain and group delay the 
system will motorboat. The solution was to add a pot that 
limits maximum ALC voltage, so it is a two-control system in 
the amp. Still, it is better to do the ALC ahead of all the 
signal path time delays, like the slower propagation through 
filters.