Low-noise oscillator for radar - phase noise floor?

Radar range resolution lives or dies on oscillator phase noise, but datasheet numbers never match the board. What phase noise floor did you actually achieve, and what dominated it - the oscillator, the PLL, or the supply?

Radar phase-noise floors on the board almost never match the oscillator datasheet; the loop and the supply usually own the number you care about.

What we see in practice:
1. Free-running OCXO / VCO hero numbers are measured in a quiet lab fixture. Once you close a PLL, in-band noise is set by reference + dividers + charge-pump / loop-filter design, not the VCO alone.
2. Supply pushing and ground bounce often dominate close-in skirts on dense boards - a clean LDO / ferrite + local bypass for the VCO / PLL rails is as important as picking a "better" part.
3. Spurs from digital clocks / switching regulators masquerade as phase noise until you look at a spectrum with markers, not only an integrated rms number.

Useful habit: publish L(f) at the offsets that matter for your waveform (e.g. 1 kHz / 10 kHz / 100 kHz / 1 MHz) *and* say whether that was free-running, locked, and on which rail. Share IF bandwidth / integration limits and whether the limit was PLL in-band or VCO far-out - that usually ends the datasheet vs board argument.

— rfforum 编辑部 / Editorial Team