L
Admin
[Technical article / FAQ, rfforum Editorial Team]
A printed dipole, loop, or other balanced antenna can have a clean S11 and still show a squinted, tilted, or unexpectedly high cross-pol pattern. Common-mode current on the feed cable or a poorly balanced balun radiates with the AUT. Return loss at the connector often cannot see that current.
1. What the balun and cable own
- S11 is a differential (or connector) match. A common-mode current on the shield can be large while the connector still looks 15-20 dB down.
- A ferrite or a quarter-wave sleeve that is slightly off frequency, or a balun that is only "pretty good" at the band edge, leaks common mode just where the pattern looks worst.
- Cable dress in the chamber changes the pattern; a remount of the AUT with a different droop that moves the beam is a feed problem until proven otherwise.
2. Freeze list before you recut the radiator
- Repeat the pattern with a ferrite on the feed, a different cable length, or a sleeve; if the squint moves or dies, stop the AUT ECO
- Balun balance: S21/S31 or a simple mixed-mode measurement, not only AUT S11
- Chamber photos of cable dress; a pattern that is not repeatable with dress is not an AUT drawing error
- Cross-pol and beam peak vs frequency: a walk that tracks balun balance is the feed
3. Habit
Never freeze a "beam squint, recut the dipole" ticket from S11 plus one pattern cut. Archive a ferrite / cable-dress A/B and a balance figure next to the pattern claim. If a ferrite on the feed restores symmetry, the radiator was never the patient.
4. Practical gate
- Add a ferrite or change cable dress; re-cut the principal planes
- Measure or at least sanity-check balun balance at the failing frequency
- Only then change arm length, flare, or matching
Discussion: on your last unexplained squint, did the fix land in the balun, cable common-mode, or a real radiator asymmetry - and do you require a dress/ferrite A/B before approving a board respin?
— rfforum 编辑部 / Editorial Team
A printed dipole, loop, or other balanced antenna can have a clean S11 and still show a squinted, tilted, or unexpectedly high cross-pol pattern. Common-mode current on the feed cable or a poorly balanced balun radiates with the AUT. Return loss at the connector often cannot see that current.
1. What the balun and cable own
- S11 is a differential (or connector) match. A common-mode current on the shield can be large while the connector still looks 15-20 dB down.
- A ferrite or a quarter-wave sleeve that is slightly off frequency, or a balun that is only "pretty good" at the band edge, leaks common mode just where the pattern looks worst.
- Cable dress in the chamber changes the pattern; a remount of the AUT with a different droop that moves the beam is a feed problem until proven otherwise.
2. Freeze list before you recut the radiator
- Repeat the pattern with a ferrite on the feed, a different cable length, or a sleeve; if the squint moves or dies, stop the AUT ECO
- Balun balance: S21/S31 or a simple mixed-mode measurement, not only AUT S11
- Chamber photos of cable dress; a pattern that is not repeatable with dress is not an AUT drawing error
- Cross-pol and beam peak vs frequency: a walk that tracks balun balance is the feed
3. Habit
Never freeze a "beam squint, recut the dipole" ticket from S11 plus one pattern cut. Archive a ferrite / cable-dress A/B and a balance figure next to the pattern claim. If a ferrite on the feed restores symmetry, the radiator was never the patient.
4. Practical gate
- Add a ferrite or change cable dress; re-cut the principal planes
- Measure or at least sanity-check balun balance at the failing frequency
- Only then change arm length, flare, or matching
Discussion: on your last unexplained squint, did the fix land in the balun, cable common-mode, or a real radiator asymmetry - and do you require a dress/ferrite A/B before approving a board respin?
— rfforum 编辑部 / Editorial Team
