[G0HZU_JMR]

Note that the circuit might not be as simplistic as a simple R and small C in the emitter. Doing my best Tom Jones impersonation "It's not unusual" for the small capacitance at the source/emitter of a transistor to look like a short circuit up at UHF. This might be due to a resonance with some wire or the equivalent.

When this happens the circuit can still oscillate up at UHF but the feedback path will be quite different. With a shorted source pin (at UHF) the Colpitts equivalent circuit we have been considering up till now ceases to function because the capacitive tap is now grounded and the feedback path is killed.

When this happens, it's the drain to gate capacitance that can become the bad guy that forms the critical feedback path. This alternative oscillator feedback is actually easier to model and understand.

Of course, it will get very muddy if you end up halfway between the two cases. The feedback path becomes very difficult to visualise and it's best to try and model it on a simulator.

The circuit below looks very similar but because the source is shorted to ground at UHF the internal feedback path is quite different. Fortunately, adding a series 100R resistor in the drain kills this mode as well.