Everybody is likely familiar with the original Armstrong
design -- tickler coil for feedback, "throttle" with
a variable capacitor RF bypass from plate to ground, etc., etc.,
etc. Keep that in mind, as this discussion progresses. If you
aren't familiar with the basic design -- grid lead detector, "tickler"
feedback and regeneration control with a variable capacitor, look
this up. Actually, Armstrong's original circuit was slightly more
complicated, but the scheme I've mentioned covers the basic design.
One needs to keep firm in mind that early receivers,
as well as transmitter designs were governed strongly by cost
and complexity considerations.
Early vacuum tubes were expensive, and parts generally
difficult to come by, and expensive as well. Also, simplicity
was a major consideration, as the more complex something is, the
lest reliable it will tend to be.
As a consequence, there were some design compromises
that we've been working to overcome over the years.
The following discussion applies to vacuum tube designs
as well as their modern cousins with bipolar and field effect
devices. They really differ only in the details. Every person
has their "pet" designs, of course, as do I. Well, here
Classic Colpitts regenerative receiver may be found: