[Elecraft] When S-Meters are Useful (WAS: When do you usefastAGC?)

[Ron D'Eau Claire]

Keith wrote:

All this AGC talk makes me think TenTec's hang AGC should have been the
ideal.  In effect it would turn the RF Gain down when a big signal came in
and keep it down until the signal went away, then would turn it back up
rather quickly.  But for some reason it just never caught on.  Maybe because
it was implemented in rigs with audio derived AGC?

---------------------------

Right on, Keith! 

Hang AGC filled a specific need. It goes way back - to the 50's at least.
Its intent is to avoid turning the gain up immediately when the signal
disappears. It keeps the gain turned down for a specific period of time.
That's the "hang" time. At the end of that period it is then turned up very
quickly. On most hang AGC systems, the "hang" time is adjustable from a few
mS to several seconds. 

It solved a tough problem in early SSB/CW AGC systems: slow attack time. A
big problem trying to create an AGC for a SSB or CW receiver was the BFO
signal in the I.F. bandpass. The AGC detector would sense that huge BFO
signal and turn down the gain accordingly, making the receiver deaf! So
early SSB/CW receivers (and the more simple receivers today, including some
of the Elecraft line) used audio-derived AGC instead of I.F. derived AGC.
That is, instead of rectifying some of the I.F. signal to produce a d-c AGC
voltage that corresponded to the strength of the signal, it rectified some
of the resulting audio signal to produce the AGC signal.

That eliminated the problem of the BFO getting into the AGC system, but it
made the attack time slow. That's because, when a signal appears, the
detector has to rectify several cycles to produce the AGC voltage. At R.F.
that signal may be hundreds of kilohertz or even higher while the audio
frequency is typically near 600 Hz. 

With audio-derived AGC it takes much, much longer to produce the d-c voltage
from the AGC rectifier. For example, it would take an AGC detector working
at 600 Hz about 200 times longer to produce the d-c voltage required than it
does when the detector is using an I.F signal at, say 100 or 200 KHz. That
delay to rectify the signal often resulted in a "pop" in the audio every
time the signal appeared as it slammed through at full volume for an instant
before the AGC voltage built up to turn the gain down. 

The first solution was to slow down the AGC action so that it only very
gradually returned the gain after a strong signal went away. That way, there
was less of a "pop" if it returned soon, but it meant that the receiver
remained deaf to weaker signals longer. Hang AGC provided a better response
by leaving the receiver desensitized for a specific period of time to
prevent another "pop" if the station was still transmitting CW or SSB, but
then quickly resumed full volume after the hang period elapsed.

Hang systems started to disappear as ways were devised to use the I.F.
signal for creating the AGC. This is most commonly done by using another
mixer and producing a new I.F. just for the AGC far removed from the BFO
frequency. The K2, for example, uses an AGC I.F. at about 200 kHz for this
purpose. That takes more circuit complexity but the fast attack time allowed
truly quick AGC without the "pops" or the delay of a hang time. 

Ron AC7AC