R
Member
I spent a week characterizing a 2.4 GHz monopole for a compact sensor node, and the ground plane turned out to matter more than the element itself. A few things I measured that surprised me.
First, shrinking the ground plane from 60 mm square to 30 mm square shifted resonance down by almost 150 MHz and dropped efficiency by about 3 dB. The current that should return through the ground plane ends up radiating from the board edge instead, which also tilts the pattern upward. If your monopole looks detuned on a small board, check the ground plane before touching the matching network.
Second, ground plane shape matters as much as size. Cutting a slot near the feed point to make room for a battery created a second resonance I did not ask for. Moving the slot 5 mm away from the feed restored a clean single resonance. Keep slots and cutouts away from the high-current region near the feed.
Third, cable effects will fool you on the bench. With a small ground plane, current flows back down the outside of the coax and your VNA sees a different antenna than the one in the product. I now use a ferrite choke right at the feed and a short length of semi-rigid coax for prototype measurements; the resonance moved 60 MHz once the choke went on, which tells you how big the error was.
Fourth, measure efficiency, not just return loss. A monopole can show a beautiful -20 dB S11 while radiating poorly because the loss is in the board, not the match. A quick Wheeler cap measurement or a simple range comparison against a known reference antenna catches this in ten minutes.
How do you handle ground plane constraints on small boards - do you extend the ground with a chassis connection, or accept the efficiency hit and budget extra link margin?
First, shrinking the ground plane from 60 mm square to 30 mm square shifted resonance down by almost 150 MHz and dropped efficiency by about 3 dB. The current that should return through the ground plane ends up radiating from the board edge instead, which also tilts the pattern upward. If your monopole looks detuned on a small board, check the ground plane before touching the matching network.
Second, ground plane shape matters as much as size. Cutting a slot near the feed point to make room for a battery created a second resonance I did not ask for. Moving the slot 5 mm away from the feed restored a clean single resonance. Keep slots and cutouts away from the high-current region near the feed.
Third, cable effects will fool you on the bench. With a small ground plane, current flows back down the outside of the coax and your VNA sees a different antenna than the one in the product. I now use a ferrite choke right at the feed and a short length of semi-rigid coax for prototype measurements; the resonance moved 60 MHz once the choke went on, which tells you how big the error was.
Fourth, measure efficiency, not just return loss. A monopole can show a beautiful -20 dB S11 while radiating poorly because the loss is in the board, not the match. A quick Wheeler cap measurement or a simple range comparison against a known reference antenna catches this in ten minutes.
How do you handle ground plane constraints on small boards - do you extend the ground with a chassis connection, or accept the efficiency hit and budget extra link margin?
