Circulator vs isolator - when do you actually need one?

PA protection - when is a circulator/isolator really necessary vs just good VSWR design? They add cost and size. What's your rule of thumb?

Circulator/isolator is insurance against *uncontrolled* load pull, not a substitute for VSWR-aware design.

Rule of thumb many PA teams use:
1. Skip it when the antenna/load is fixed, well-matched across the band, and you have already proven |Γload| stays inside the PA’s safe contour (with cable flex, ice, people walking by). A good output match + VSWR foldback / current limit often wins on cost and size.
2. Put one in when any of these are true: antenna can go open/short (connector fail, ice, deploy/retract), multi-antenna or switched paths with unknown Γ, high-Q filter after the PA that looks reactive off-tune, or GaN/LDMOS that walks into instability or thermal runaway under high |Γ|.
3. Prefer a *circulator + load* (true isolator) over a 3-port circulator with an open third port. Spec the isolator for the PA’s peak reverse power and the duty cycle you actually see in fault, not only CW datasheet.

Cost/size middle path: directional coupler + detector for VSWR foldback first; add a ferrite isolator only on the SKUs that fail the open-load soak. Open question: is your threat a hard open at the connector, or soft mismatch from a filter/antenna that still passes DC continuity?

— rfforum 编辑部 / Editorial Team