Polarization is the one antenna specification that changes the shape of the read zone rather than its size. Choose wrong and the deployment will read 98 % of tags in one orientation and 60 % in another, which is worse than a uniformly shorter range because it looks like a random failure.

The trade is simple to state: a circular antenna is forgiving of tag orientation and costs about 3 dB of link budget; a linear antenna gives that 3 dB back and creates nulls when the tag is cross-polarized.

What is polarization, physically?

The electric field of the reader's carrier has an orientation. A tag antenna is a linear structure, so it responds strongly when its axis aligns with that field and hardly at all when it is perpendicular.

Because nobody controls how a carton is placed on a pallet, a tag arrives in an arbitrary orientation. Fixed readers must therefore decide whether to design for the worst orientation or to exploit the best one.

Linear: the field oscillates in one plane. Best case is +3 dB over circular, worst case is a deep null.
Circular: the field rotates. Constant response regardless of tag rotation, at the cost of 3 dB.
Cross-polarization rejection: two linear antennas at 90 degrees barely hear each other, which is useful for isolation.

What does linear polarization actually buy you?

Linear gives back the 3 dB that circular spends, which translates directly into range — roughly 40 % more distance in free space, and often more in practice because the margin also covers cable loss and tag inefficiency.

It also gives control. Because a linear antenna only illuminates one plane, it is the right tool when you must read a specific face of an asset and avoid reading the adjacent rack. Shelf-edge antennas and narrow conveyors are the classic examples.

Range: typically 15–40 % more than circular at equal reader power.
Orientation: the tag must be roughly aligned within 45–60 degrees of the antenna axis.
Best use: conveyor tunnels, shelf edges, single-face asset reads, controlled orientation.

What does circular polarization cost, and when is it worth it?

Circular polarization is produced by feeding a patch or crossed dipole with two signals 90 degrees out of phase. The field rotates once per RF cycle, so a linear tag sees a constant amplitude regardless of its rotation.

That constant amplitude comes from splitting the power between two orthogonal components: each carries half, so the tag receives 3 dB less than a perfectly aligned linear case.

In exchange, read rate becomes insensitive to how the carton landed. In a dock door, where pallets arrive in every orientation, that trade is almost always worth it.

Typical axial ratio: 1–3 dB on good antennas, 5 dB or worse on cheap ones.
Best use: portals, dock doors, overhead readers, handhelds, any uncontrolled orientation.
Penalty: 3 dB, plus a larger physical antenna for the same gain.

Which antenna belongs in which read zone?

Match the antenna to how much control you have over tag orientation.

Dock door and portal: circular, always. Orientation is uncontrolled and throughput matters.
Conveyor tunnel: linear, because the tag face is known and every decibel converts to speed.
Shelf edge: linear, pointed across the shelf, so the neighbouring bay is not illuminated.
Overhead inventory reader: circular, because garments and cartons rotate freely.
Handheld: circular, because the operator cannot align the tag.

A common mistake is using circular everywhere "for safety". In a conveyor tunnel, that decision costs range you already paid for and reduces the speed at which the line can run.

What is axial ratio, and why is cheap circular not circular?

Axial ratio is the ratio between the major and minor axes of the polarization ellipse. A perfect circular antenna has 0 dB; a practical one is 1–3 dB; a poor one drifting toward 6 dB is closer to elliptical and behaves like a mismatched linear antenna.

Cheap antennas cut cost in the feed network, and that is exactly where axial ratio is determined. A 6 dB axial ratio can lose 1.5–2 dB depending on tag orientation, which erodes the very benefit you bought circular for.

Ask for the axial ratio across the band, not at one frequency.
Verify on a real tag population, in several orientations, not on a single test tag.
If the datasheet omits axial ratio, assume it is poor.

Can you mix circular and linear in one installation?

Yes, and it is often the best design. A portal with circular antennas at the sides and a linear antenna above the belt can cover both the uncontrolled pallet face and the known top face.

Two rules keep mixing safe. First, never place a circular and a linear antenna pointing at the same tags with equal power — the linear one will dominate and recreate the orientation sensitivity you were trying to remove. Second, balance by zone: each antenna owns a defined region.

Assign each antenna a role and a power setting in the controller.
Measure per antenna, and confirm that the linear one is not masking the circular one.
Keep neighbouring portals on different channels to avoid cross-reading.

How do you verify polarization in the field?

Take one tag, rotate it through 90 degrees in 15-degree steps, and record RSSI at each step. The signature tells you what you actually installed.

Flat response within ±2 dB: genuine circular, working as intended.
Response varying 10 dB or more: linear, or a circular with a bad axial ratio.
Deep null at 90 degrees: linear, and the tag population will hit that null regularly.

Repeat with the tag at the far edge of the zone and at the worst height. Polarization problems are worst where the link budget is thinnest, which is exactly where you should be measuring.

Bottom line

Circular buys forgiveness for 3 dB; linear buys range and control for orientation discipline. Portals, overhead readers and handhelds want circular; conveyors, shelf edges and fixed-face reads want linear. Verify axial ratio across the band and confirm the pattern with a rotation test, because the datasheet rarely tells the whole story.