Short answer: you can make a cavity “as big as you like,” but past a sensible size you don’t gain performance—and you often make things worse (more spurious modes, weight, cost, drift). A 50 cm length at 1420 MHz is well into “no real advantage.”
Why size helps—up to a point
- Unloaded Q rises with physical size because RF currents spread over more area and conductor loss per unit length falls. This lowers insertion loss and steepens skirts for the same coupling.
- But increasing size also lowers the spacing to unwanted modes (axial TE/TM modes in a long can, higher-order TEM modes in long coaxial resonators) and makes mechanical/tuning stability harder.
Two common 1.42 GHz cavity types
- Cylindrical TM010 cavity (pill-box)
Resonant frequency depends mainly on radius, not length:
f{TM010} =approx 2.405 x c} / (2 x pi x a)
Where a is the radius.
At 1420.405 MHz this gives radius (a) of approx 80.8 mm) (≈ 161.6 mm diameter). The height can be quite short; making it tall (e.g., 50 cm) doesn’t change the TM010 frequency but does introduce closely-spaced axial modes (TM({011}), TE({111}), …) that complicate tuning and can spoil stop-band behavior.
Rule of thumb: keep height on the order of a fraction of the radius (e.g., 30–120 mm) unless you have a specific coupling/packaging need.
- Quarter-wave coaxial (combline/interdigital) resonator
Frequency is set by electrical length ~ λ/4; in air:
Wavelength approx. c / f which is approx. 0.211m. This means one quarter wavelength is approx 53mm.
End capacitance and re-entrant designs adjust the physical length a bit, but we’re still talking a few cm, not tens of cm. Making the resonator long (50 cm ≈ 2.4 λ) doesn’t improve Q meaningfully; it just invites higher-order TEM modes and mechanical headaches.
What you can expect at 1420 MHz
- Q and loss: A well-built copper or silver-plated TM010 cavity of the diameters above can reach unloaded Q in the low-five-figures; combline resonators are usually lower (few-thousand) but easier to cascade compactly. Going dramatically longer than needed does not buy you proportionally higher Q.
- Spurious modes: Long cavities pack more modes near your passband. For astronomy (weak-signal, strong RFI nearby), mode-free stop-band matters; keep the axial dimension modest and use proper mode-suppression features (chamfers, absorbers away from the high-Q region, careful probe/loop placement).
- Thermal drift: Fractional drift is set by material CTE; making it larger doesn’t change ppm/K, but you now have bigger absolute motion and more mechanical sensitivity. Better results come from temperature control, Invar/CTE-compensated posts, and low-loss plating, not from making it long.
Practical guidance
- For a TM010 single-resonator preselector at 1420 MHz: target Ø ~160 mm, height ~40–100 mm, critically coupled to reach your desired bandwidth; silver-plate the RF surfaces; keep joints RF-tight.
- For a multi-pole filter with compact size: use combline/interdigital with ~50–80 mm resonators, stout inner/outer conductors to keep conductor loss low, and absorptive lining in spurious hot-spots if needed.
- If you truly need lower loss, prioritize surface finish, plating, corners/radii, and coupling design before increasing bulk dimensions.
Bottom line: At 1420 MHz, a 50 cm-long cavity offers no real benefit and brings real risks. Increase diameter (for TM010) or conductor cross-section (for combline) modestly and focus on plating, coupling, and temperature control for meaningful performance gains.