Duplexer isolation for FDD small cells - how much is enough?

FDD small cells with tight Tx-Rx spacing struggle with receiver desense. What Tx-Rx isolation did you design for, and how much margin did you keep over the 3GPP minimum?

FDD small-cell duplexers: "enough" isolation is whatever keeps the victim RX noise floor and blocker budget inside the desense mask at your max TX power and closest channel spacing.

A workable framing:
1. Budget RX noise rise from TX noise in the RX band (filter rejection + PA noise) separately from RX compression / IM from residual TX leakage.
2. Board / cavity leakage and antenna isolation often dominate once the ceramic/SAW/BAW duplex filter is "good enough" on paper — measure the assembled unit, not only the filter datasheet.
3. Typical cellular small-cell designs chase tens of dB of filter rejection plus careful layout; exact dB depends on duplex gap, PA ACLR, and LNA IIP3.
4. If isolation is short, fix TX noise and shielding before buying a more exotic filter — filter Q alone rarely rescues a leaky PA–LNA layout.

Share duplex gap, TX power, and whether the front-end is discrete or a PAMID-style module; that usually pins the isolation number people quote.

— rfforum 编辑部 / Editorial Team