[Lowfer] Lowfer KE

[J D]

Kasey, it would appear from your schematic that a few components may be 
missing, thereby preventing your oscillator from being as stable as it 
needs to be. This same situation is found in a commercially made 
sorta-kit rig for 22 meters, and beacons using it there can often be 
identified by their tendency to wander or even outright jump around in 
frequency from day to day.

A Pierce oscillator needs capacitance from both legs of the crystal to 
ground as part of a phase shift network. There is stray capacitance 
present, of course, without which it would not oscillate at all; but 
that is neither of sufficient magnitude nor stability to ensure reliable 
oscillation at the design frequency of the crystal. A good starting 
value for capacitors for a typical 3 MHz parallel resonant crystal would 
be 27 pF each.

In addition, there needs to be a series resistor between pin 10 and the 
crystal. It serves three functions that each affect frequency stability 
to a certain extent: it isolates the crystal somewhat from the complex 
impedance of the gate output, it limits current and thereby power 
dissipation in the crystal, and it is part of the phase delay network. 
As a rough first approximation, I'd suggest 1 K, maybe as high as 2.2 K.

Finally, the shunt resistance across the crystal could be a little 
higher in value for 3 MHz operation, such as 3.3 M to 4.7 M. It's not 
terribly critical, but it is preferable to operate it with as high a 
value as will allow for reliable startup. Lowering the operating Q of 
the crystal with a low parallel resistance could allow a crystal to jump 
too easily between primary resonance and any secondary resonance modes 
it may have.

These comments are some basic rule of thumb guidance. For more detailed 
treatment of Pierce oscillators implemented with logic gates, I believe 
Crystek still offers this article:

https://www.crystek.com/documents/appnotes/pierce-gateintroduction.pdf

John Davis