Reader Power vs Read Range: Where the Decibels Actually Go

Turn up the power and the tags will be read further away. It sounds obvious, and in a passive UHF system it is only partly true. Doubling reader power buys 3 dB, which buys roughly 40 % more range in free space — and often much less in a warehouse, where the limit is usually the tag's wake-up threshold, the orientation, or the regulator rather than the amplifier.

Understanding which of those three is binding is what prevents both under-powered designs and illegal ones.

What does doubling the power actually buy?

Radio power falls with the square of distance in free space. In decibels, doubling the distance costs about 6 dB; doubling the transmitted power adds 3 dB, which extends range by a factor of about 1.41.

That is the entire arithmetic, and it explains why the last few metres are so expensive. Going from 4 metres to 8 metres requires four times the power, not two — and that is before real-world losses.

• 2× power = +3 dB = 1.41× distance.
• 4× power = +6 dB = 2× distance.
• 10× power = +10 dB = 3.16× distance.

In a warehouse the effective exponent is worse than free space because of multipath, absorption and polarization loss, so 4× power often buys only 1.5× range.

Why is the tag's wake-up threshold the real limit?

A passive tag needs roughly 15–30 microwatts at its rectifier to power up. That number, not the reader's output, is what sets the far edge of the zone: once the received power falls below the threshold, the tag is silent no matter how loudly the reader shouts.

Tag sensitivity varies widely. A modern Gen2 chip with a well-matched antenna wakes at around -20 to -22 dBm; an older or poorly matched inlay may need -14 dBm. That 6–8 dB difference is equivalent to a 4–6× change in reader power.

• Check the chip's read sensitivity, not the marketing "read range".
• A 3 dB better tag is worth 2× the reader power, at a fraction of the energy cost.
• Cheap tags with narrow bandwidth lose sensitivity at the band edges, which shows up as region-specific failures.

What are the legal ceilings, and why do they differ?

Every region caps EIRP, and the cap is not negotiable. Europe: 2 W ERP (3.3 W EIRP) in the 865–868 MHz band. United States: 4 W EIRP in 902–928 MHz. China: 2 W EIRP in 920–925 MHz. Japan: 1 W EIRP typical in 916–921 MHz.

The differences mean a reader tuned for the US market will be over the limit in Europe and Japan. Good readers expose a region setting that caps output and applies the right frequency plan; cheap readers let you type a number and hope.

• Always configure the region before the first site survey.
• A 1 W EIRP difference is about 6 dB of link budget — enough to change the number of antennas required.
• Document the configured EIRP per reader; auditors and neighbours both ask.

Where else do the decibels disappear?

Between the amplifier and the tag, several losses are unavoidable and one is usually avoidable.

Cable loss is the avoidable one: 3 dB per 10 metres of thin coax at 900 MHz is common, and a cheap connector adds 1–2 dB. Antenna gain partially compensates: a 9 dBi panel versus a 6 dBi patch is 3 dB, roughly doubling the range for the same reader power.

Polarization mismatch averages 3 dB and can reach 20 dB when a linear tag sits cross-polarized to a linear antenna. Tag antenna gain of 1.5–2.2 dBi and rectifier efficiency of 20–40 % complete the chain.

• Cable: 1–3 dB typical, 6 dB in a badly planned installation.
• Connectors and lightning arrestors: 0.5–2 dB.
• Polarization: 3 dB average, 20 dB worst case.
• Tag efficiency: the largest non-obvious loss.

Does more power ever make things worse?

Yes, in three situations. In dense deployments, extra power increases the interference radius of every reader, which raises the noise floor and can reduce total throughput even as individual reads get stronger. Dense-reader mode exists precisely because of this trade.

Second, over-powering near a metal structure can create strong reflections and nulls where the tag sits in a cancellation point, so the zone becomes patchy rather than larger.

Third, a reader at maximum power runs hotter, and thermal drift detunes its own front end — a fixed reader that reads worse after two hours is often simply too hot.

• Prefer more antennas at moderate power over fewer antennas at maximum power.
• Leave 20–30 % power headroom for tuning rather than running at the cap.
• Monitor temperature and RSSI distribution over a shift, not at commissioning only.

How should power be set in practice?

Set power per antenna, then verify with a rotation test at the far edge of the intended zone.

Start at 60–70 % of the legal cap, measure the read rate on the real tag population in the worst orientation, then increase only if the shortfall is caused by link budget rather than by geometry.

• Measure at the zone edge, not at the antenna.
• Record read rate at three power levels and pick the one that meets the process requirement with margin.
• Re-check after any change to tags, pallet build or antenna mounting.

Bottom line

Reader power is logarithmic, capped by law, and often not the limiting factor. Fix the tag first, then the antenna, then the cabling, and only then raise the watts — within the legal ceiling. A 6 dB gain from a better tag or a shorter cable is worth more than quadrupling the amplifier.