[Lowfer] Beacon Schematic

[J D]

The Rakon note is interesting and goes into a bit more detail than the 
Crystek paper I referenced earlier, but it also has some technical flaws 
(apart from reckless disregard of SI conventions for writing units, and 
a few parts appearing to be edited by someone who is not a native 
speaker of either American or English English; but that's just my 
frustrated-English-major background coming through :). A couple of examples:

One can express gain as either a voltage ratio or a power ratio in dB if 
labelled accordingly, but "dBV" signifies the ratio of some measured 
voltage to a reference value of 1 volt.  That creates a confusing 
ambiguity at best in Figure 4.

It's also not quite accurate to state that the reason for preferring 
unbuffered inverting gates is primarily because of "the associated gain 
of many thousands" in the buffer stages. True, a higher open-loop gain 
can make unity-gain compensation trickier in a feedback loop, but the 
major problem in this case is the propagation delay associated with the 
two additional inverters that make up the buffer. Nowadays, that's less 
of a problem, and it's entirely reasonable to build stable crystal 
oscillators up to 8 or 10 MHz without any fuss using buffered devices 
like the 4060--so long as the feedback and series resistors are employed 
appropriately. It's not entirely out of the question, even, to do 14-17 
MHz with buffered gates, but more care should be taken there in 
selecting the gate type, crystal parameters, and loop component values. 
The higher in frequency you go, the more you'll want to do the math. 
(Lack of attention to those details may be main the reason why many 
Black Cat Systems 22 meter "kits" tend to wander over hundreds of Hz 
from day to day...when the oscillator bothers to start at all, that is.)