L
Admin
[Technical article / FAQ, rfforum Editorial Team]
A metal lid or package shield is drawn as the isolation fix. On the board, sparse stitching vias around the RFIC and under the can turn the "shielded" compartment into a leaky cavity: energy couples under the fence before the lid ever matters.
1. What the fence actually does
- Via pitch and row count set the stopband of the fence; a single sparse ring can look fine at low GHz and open up at the product band
- Gaps at connector cutouts, crystal keep-outs, and DC feed-throughs dominate leakage more often than the package itself
- A can that "passes" with copper tape over the seams can fail when production lid torque and via plating differ from the prototype
2. Freeze list before you blame the die or the lid
- Via pitch / row count around the RFIC and under the can wall, on the same sheet as isolation or spur claims
- Whether cutouts were modeled or measured with the real connector / crystal footprint
- Lid contact (torque, gasket, plating) and whether the traveler used the same lid as production
- Which port was driven when isolation was claimed (near-field probe vs conducted neighbor)
3. Habit
Never freeze an "RFIC shielded" claim on the package drawing alone. Archive fence geometry and cutout notes next to every isolation / spur traveler.
Discussion: on your last RFIC isolation miss, did the fix land in denser stitching, a lid/gasket change, or a layout cutout — and do you require via pitch on the same sheet as the isolation number?
— rfforum 编辑部 / Editorial Team
A metal lid or package shield is drawn as the isolation fix. On the board, sparse stitching vias around the RFIC and under the can turn the "shielded" compartment into a leaky cavity: energy couples under the fence before the lid ever matters.
1. What the fence actually does
- Via pitch and row count set the stopband of the fence; a single sparse ring can look fine at low GHz and open up at the product band
- Gaps at connector cutouts, crystal keep-outs, and DC feed-throughs dominate leakage more often than the package itself
- A can that "passes" with copper tape over the seams can fail when production lid torque and via plating differ from the prototype
2. Freeze list before you blame the die or the lid
- Via pitch / row count around the RFIC and under the can wall, on the same sheet as isolation or spur claims
- Whether cutouts were modeled or measured with the real connector / crystal footprint
- Lid contact (torque, gasket, plating) and whether the traveler used the same lid as production
- Which port was driven when isolation was claimed (near-field probe vs conducted neighbor)
3. Habit
Never freeze an "RFIC shielded" claim on the package drawing alone. Archive fence geometry and cutout notes next to every isolation / spur traveler.
Discussion: on your last RFIC isolation miss, did the fix land in denser stitching, a lid/gasket change, or a layout cutout — and do you require via pitch on the same sheet as the isolation number?
— rfforum 编辑部 / Editorial Team
