One often sees a small resistor (e.g. 22 - 100 ohms) inserted in series with the base in VHF/UHF/Microwave common base amplifier circuits. I believe this aids stability but I don't understand how it works. I'd be grateful if someone could explain.
I'm building an 80 MHz crystal oscillator based on  and I had a spurious ~ 1.3 GHz parasitic oscillation which went away when I increase the base series resistance of the top cascode transistor from 22 to 100 ohms. I haven't tried values in-between yet. I'm using BFM520 instead of the obsolete and unobtainable device specified in .
I'm also curious why one would need a transistor with such a high fT (several GHz) for a 100 MHz oscillator. Is it because they come with very low junction capacitances and low noise figures, or do we actually need the fT as well?
Transistors with a capacitive loaded emitter will oscillate very quickly with some base inductance. The feedback path is via the BE diffusion capacitance. I use this property frequently in oscillator design when design frequency is not far below ft.
If you have a simulation program, "make" a circuit with grounded collector. Drive the base with an AC current source and "measure" the voltage. When you do an AC analysis in PSPICE, you can select 1A current. The voltage across the current source equals the input impedance of the transistor. When the emitter is grounded for AC, you will measure dissipative impedance.
Now, play with the emitter capacitance (provide a DC path via a current source, so the circuit has DC bias). You will notice negative real part in input impedance. Just a small inductance (lead inductance) can be sufficient to make a nice oscillator. Adding some base resistance kills the negative resistance and avoids parasitic oscillation due to parasitic inductance in the base circuit.
When you make a cascode or common base circuit, there can be frequency ranges where the emitter sees capacitive impedance. To avoid parasitic oscillation, the base resistor may help. Ferrite does also a good job (as this doesn't affect noise behavior mostly). The difficulty with ferrite is to know what you have, so the resistor is a good alternative.
Regarding ft. High ft will not always result in low noise. Oscillator phase noise close to the carrier is caused by LF noise in the transistor (and current carrying resistors).
Using high ft transistors can be useful in isolating amplifiers (to avoid frequency drift under varying load). Low Cce and Ccb will result in less feedback in case of a common base amplifier (cascode circuit).
Also high ft transistors behave more ideally, so design can be easier (except for parasitic oscillation).