Cost/Benefits of an RF Cavity vs Quartz Oscillator

[Twigg]

I just encountered an RF cavity in the wild for the first time. It was used as the frequency reference in the Agilent 8640B RF signal generator, which I believe dates back to the 70's. Were quartz oscillators not an option back then? Or were they worse in stability back then? I'm curious about the cost/benefit analysis from that period, if anyone knows. Thanks!

 

[tech99]

Quartz crystals cannot be made for fundamental frequencies above approx 24 MHz. Operation at overtone frequencies allows operation up to about 250 MHz, but at low power. Of course, harmonics of the crystal oscillator can be used at higher frequencies. Cavity resonators become larger as the frequency goes down, so are impractical inside test equipment, below, say, 1000 MHz. Transmission line resonators are good for the VHF range where the dimensions are practicable, but at lower freqeuncies the line losses increase making them less attractive.

 

[Baluncore]

A tuned RF cavity can be used to select one harmonic of a crystal oscillator. For example, a 2500 MHz high-Q cavity can select the 25th harmonic of 100 MHz, generated from a lower frequency crystal oscillator. That produces a very low phase noise signal that can be used to down-convert microwave signals. PLL synthesizers have higher phase noise and so are not suitable for high quality instruments such as microwave spectrum analysers.
The problem with narrowband high-Q RF cavities is that they must be physically rigid and temperature stabilised.

 

[Baluncore]

The RF source for the HP8640B is a 256 to 512 MHz mechanically tuned cavity, with electronic varactor control to trim the frequency and to introduce FM.
Frequency accuracy comes from the inbuilt frequency meter that is based on an internal 5 MHz crystal oscillator.

The HP service manual is helpful.

The cavity is a foreshortened type which is essentially a length of coaxial transmission line with a short at one end and a capacitive load at the other. The shorted transmission line is less than 1/4 wavelength long at the frequency of oscillation and its impedance is inductive. The cavity resonates at the frequency at which the inductive reactance of the transmission line equals the capacitive reactance of the load capacitor.

The resonant frequency is varied by changing the length of the cavity (a secondary effect) and by changing the load capacitance. The varactor diodes are in parallel with the main load capacitance. The cavity is mechanically fine tuned by protruding a small metal slug into the cavity. Signal is coupled out of the cavity into two buffer amplifiers by loops which protrude into the cavity.

 

[DaveE]

One advantage of RF cavities is they are easily tunable.

Anyway, there is a ton of info about the 8640B. Here's an interesting HP Journal article about the design.
https://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1973-02.pdf