Circulator isolation that collapses when the load port VSWR is only 1.5:1

Started by LunaWei · 2026-11-16 01:49 UTC · 19 views · 2 replies
L Admin
[Technical article / FAQ, rfforum Editorial Team]

A circulator datasheet isolation assumes a well-matched load port. In a real chain, antenna VSWR of 1.5:1 (or a poorly terminated dump) reflects power back into the junction and the "isolation" you measure at the third port collapses — even when the ferrite part is fine.

1. What the load port owns
- Isolation is not a three-port constant; it is conditional on the match at the other ports. Reflected power from a mistuned antenna re-enters the junction and appears as leakage toward the LNA or PA sense path.
- A lab termination that is excellent makes the part look hero. The same unit on a 1.5:1 antenna looks broken.
- Temperature and magnet bias shift the match too; a cold bench pass can fail after soak on the tower.

2. Freeze list before you swap the circulator
- Load-port VSWR / return loss at the exact frequency of the isolation claim (antenna or dump, not the cal load alone)
- Isolation vs a swept load mismatch (slide screw or known mismatch) — if isolation tracks load Gamma, stop the part ECO
- Cable and adapter stack between circulator and antenna; a bad jumper invents "circulator fail"
- Case temperature under the same RF drive as the field condition

3. Habit
Never freeze isolation from a 50 Ω bench load alone when the product sees an antenna. Archive load VSWR on the same sheet as isolation. If a better match restores isolation, the ferrite was never the patient.

4. Practical five-minute gate
- Measure antenna / dump RL at the circulator plane
- Substitute a good load; note isolation recovery
- Write isolation @ stated load VSWR on the traveler, not "datasheet isolation"

Discussion: on your last isolation miss, did the fix land in antenna match, jumper, or a real circulator — and do you forbid isolation claims without the load VSWR on the same line?

— rfforum 编辑部 / Editorial Team
L Admin
Addendum — isolation belongs next to the load VSWR, not next to the datasheet headline. A unit that looks heroic into a lab termination and collapses at 1.5:1 is doing what the junction physics said. A cheap gate: measure antenna or dump return loss at the circulator plane, then substitute a good load; if isolation recovers, stop the part ECO.

A swept mismatch (slide screw or known Gamma) is even better — if isolation tracks load Gamma, the ferrite was never the patient. Write the jumper stack and case temperature on that same line; a bad jumper or a cold-bench pass that fails after soak invents "circulator fail".

Open: on your last isolation miss, did the fix land in antenna match, jumper, or a real circulator — and do you forbid isolation claims without the load VSWR on the same line?

— rfforum 编辑部 / Editorial Team
P Member
Good point above. One addition: conclusions can differ a lot under different test conditions.
广告 AdRenesas 瑞萨微控制器芯片产品照片Renesas Electronics微控制器、模拟与射频/时钟方案 · MCUs, analog, RF and timing访问官网 →免费展示 · 资料来自公开信息 · 如需撤下请联系 ads@rfforum.com
广告 Ad广告位招商 / Advertise hereads@rfforum.com · 刊例 Rate cardT2 · 728×90
Log in or register to reply.