Free tool
Surface test
The same tag on cardboard, glass, a bottle of water and a steel rack. See what each surface does to the read, and what a spacer or an on-metal tag gets back.
Change the surface
Power at the chip
Reply at the reader
- Range on cardboard
- Range on this surface
- Lost to the surface
The surface figures are typical published values, not our measurements. We're replacing them with bench readings, same tag and reader, surface swapped. Until then, treat the ranges as the shape of the answer, not the number.
Free space otherwise: no floor, no walls, no stock. A real room is usually 30-40% shorter again. Wake the tag has the same sums without the surface.
How it works
What you’re looking at
The tag doesn’t change. The surface under it does, and the surface decides how much antenna the tag has left.
On cardboard it has all of it. Press it flat on steel and the metal makes a mirror image of the antenna with the current running the other way. The two cancel, and right at the surface the field is close to nothing. The chip gets about one per cent of what it got on card. That’s why a tag on a steel rack goes quiet.
Water is different. It doesn’t cancel the wave, it absorbs it. A bottle of shampoo, a crate of meat, a person walking past: all mostly water, all soaking up the signal at 866 MHz.
A spacer helps. A few millimetres of foam between label and steel lets the antenna work again, by an amount that depends heavily on the stand-off. A tag made for metal does better: it’s a patch antenna that uses the steel as its ground plane, so it wants the metal there.
The figures are placeholders
The gain penalty for each surface is a typical published figure, not a measurement of ours. We’re replacing each one with a bench reading: same tag, same reader, same distance, surface swapped. The source line on each surface says which it is. Until then, the ranges here are the shape of the answer, not the number.
What it assumes
Free space apart from the surface: no floor, no walls, no stock. Real rooms, with floors, walls and stock, usually do worse.
The reader puts 1 W into 1 dB of cable and a 6 dBi circular antenna, at 866 MHz. That is 35 dBm EIRP, a hair under India’s limit of 2 W ERP (35.15 dBm EIRP). The tag is an NXP UCODE 9 on a half-wave dipole; the reader is an Impinj E710 in dense-reader mode. These are the same numbers as Wake the tag, which lets you move the tag without the surface in the way, and the range calculator lets you change all of them.
Where the numbers come from
Chip and reader sensitivities are NXP’s and Impinj’s own figures, listed on the range calculator. The surface penalties are in src/config/rfid-surfaces.ts with their sources, and every one is marked as a placeholder until the bench says otherwise.