Which frequency band belongs to which application, and why

01 / FIELD NOTE

Keep the decision tied to the operating context.

Every RFID application picks its frequency band before it picks anything else, and the choice is not fashion: range, power budget, and the way water and metal bend the signal sort the work in advance. So the useful form of the question "which band for which job" is really "what does this band's physics give it that the job needs". A livestock ear tag, a door card, a case of goods on a pallet and a container crossing the ocean are read at very different distances, in very different materials, with very different tolerances for cost — and each band answers one of those constraints precisely because of how its read works.

RFID splits into two coupling regimes, and the split explains most of the grouping. Low and high frequency are inductive: the reader's coil generates a magnetic field and the tag's coil harvests it, so coil geometry and spacing set the reach, not transmitter power — beyond a few tens of centimeters, more current buys almost nothing. The same physics makes the read nearly immune to water, since a magnetic field passes through it freely, while large metal plates short it out. Ultra-high frequency is radiative: the tag scavenges the reader's wave and backscatters a response, so range stretches to meters, and water and metal become its enemies — water absorbs and metal mirrors the traveling wave.

Low frequency at 125 kHz takes the moisture-tolerant, contact-range jobs. An injected chip under a pet's skin sits in aqueous tissue, yet the magnetic field couples through it with little loss, and an ear tag on a wet animal still reads at a handheld reader's distance. The same short-reach behavior is what makes the car immobilizer work: the coil in the key and the ring around the ignition only need to couple across centimeters when the key turns, so no battery is required and the system cannot be interrogated from outside the car — the tiny range is the anti-theft feature.

High frequency at 13.56 MHz keeps the inductive ceiling and adds a faster clock, so it trades the same short range for more data per transaction and cheaper cards, and it stays comfortable around damp materials. That is the profile behind the access card on a door reader, book labels on library shelves, and linen in a laundry: read contact-close, tolerate moisture, cost little. Near-field communication runs on the same physical layer but adds a standardized protocol stack with a secure element, and its deliberately short reach is a design decision — payment and wallet access want the user to intend the proximity.

Ultra-high frequency, the 860 to 928 MHz band, is the supply-chain workhorse because far-field backscatter changes what reading means. A tag need not be presented; it answers from meters away, cases and pallets can be swept in bulk without line of sight, and inlays are cheap at volume, which makes disposable slap-and-ship labelling economic. Dry goods cooperate with the physics: cardboard, plastic and packed items neither absorb nor mirror the wave strongly. The same physics sets the limit — cartons of liquid and anything behind a metal frame or steel rack eat into the read, which is why UHF settles where the goods are dry and the reads are at distance.

At 433 MHz the defining move is the battery. An active tag broadcasts its own signal instead of reflecting the reader's, so the range jumps to tens of meters, the link works through container walls and stacked loads, and the lower frequency propagates more kindly around steel ribs and damp interiors than the higher supply-chain band would. The battery also hosts sensing: a seal sensor or a temperature logger can watch the shipment because the tag does not depend on reader energy. That cost — battery, housing, duty cycle — confines the design to high-value cargo, and decades of military logistics runs, led by the U.S. DoD, have made the band's long-range behavior the benchmark for container and intermodal tracking.

Microwave at 2.45 GHz is mostly active too, and it sits in the same licence-free band as Wi-Fi, which is precisely its trick: the existing access-point infrastructure can listen, so a real-time location system rides on the Wi-Fi network and measures timing across base stations instead of installing a dedicated reader grid. The higher carrier also carries more bits per second, which matters for the band's other classic job, toll collection — the active tag in the vehicle has only the few meters of a lane to exchange identity and toll data at speed. The vehicle body's metal is a mirror, so the mount is planned around the windshield rather than behind body panels.

Ultra-wideband is the odd band out because it barely concerns range or power at all. It transmits extremely short pulses across a wide slice of spectrum around 5 GHz, and the pulse width is the whole game: a sub-nanosecond pulse lets a receiver measure time of arrival closely enough to fix a tag's position to centimeters, and the wide spectrum supports high data rates. That accuracy costs reach — the power limits that keep the band unlicensed confine it to tens of meters indoors — so the applications are location rather than passage: equipment finding in workshops and rooms, and real-time location systems where the answer demanded is "where", not "that".

The groupings above are what the field settled into, not what the physics forbids. Engineers routinely push bands outside their usual homes — UHF into laundries, HF onto metal with ferrite spacers — and regional frequency allocation rules differ, so a band that is freely available in one territory may be contested in another. The cheat-sheet answer is a starting point: name the application, and the band that matches its distance, its material, and its price is usually the one the industry already converged on — but the fit is a judgement, not a verdict, and the honest next question is always to measure the read in the actual installation.