L
Admin
[Technical article, rfforum Editorial Team]
A single patch that looks fine on a large ground plane can misbehave badly once it sits in a 2×2 or 4×4. Mutual coupling changes the active reflection coefficient with scan angle and can open a blind spot long before the passive S11 of the isolated element looks wrong.
1. Passive S11 is not active Γ
The isolated-element match tells you about the feed and the local cavity. Under scan, neighbouring elements inject currents that shift the impedance the PA or LNA actually sees. Quote active impedance (or scanned S-parameters) for the scan volume you care about.
2. Embedded element pattern (EEP)
Measure or simulate one driven element with neighbours terminated in the system impedance. The EEP already folds in mutual coupling; array factor on top of a free-space single-patch pattern will over-predict scan performance.
3. What fails first on small arrays
- Edge elements vs centre elements: different active Γ
- E-plane scan usually hurts harder than H-plane for patches
- Surface-wave / substrate modes that only show once the lattice is dense
- Ground-plane and feed-network common-mode that the single-patch fixture never excited
4. Practical bring-up
- Freeze a thru-calibrated active-impedance check at broadside and at the max scan angle before PA matching
- Compare EEP cuts against the isolated pattern; large disagreement means coupling, not a bad single match
- If you retune the feed for broadside only, re-check the scan edge — you may have traded one problem for another
Discussion: on your last small patch array, which failed first under scan — active match, EEP null, or feed-network imbalance — and did you fix it in lattice spacing, substrate, or matching?
— rfforum 编辑部 / Editorial Team
A single patch that looks fine on a large ground plane can misbehave badly once it sits in a 2×2 or 4×4. Mutual coupling changes the active reflection coefficient with scan angle and can open a blind spot long before the passive S11 of the isolated element looks wrong.
1. Passive S11 is not active Γ
The isolated-element match tells you about the feed and the local cavity. Under scan, neighbouring elements inject currents that shift the impedance the PA or LNA actually sees. Quote active impedance (or scanned S-parameters) for the scan volume you care about.
2. Embedded element pattern (EEP)
Measure or simulate one driven element with neighbours terminated in the system impedance. The EEP already folds in mutual coupling; array factor on top of a free-space single-patch pattern will over-predict scan performance.
3. What fails first on small arrays
- Edge elements vs centre elements: different active Γ
- E-plane scan usually hurts harder than H-plane for patches
- Surface-wave / substrate modes that only show once the lattice is dense
- Ground-plane and feed-network common-mode that the single-patch fixture never excited
4. Practical bring-up
- Freeze a thru-calibrated active-impedance check at broadside and at the max scan angle before PA matching
- Compare EEP cuts against the isolated pattern; large disagreement means coupling, not a bad single match
- If you retune the feed for broadside only, re-check the scan edge — you may have traded one problem for another
Discussion: on your last small patch array, which failed first under scan — active match, EEP null, or feed-network imbalance — and did you fix it in lattice spacing, substrate, or matching?
— rfforum 编辑部 / Editorial Team
