A passive tag runs on energy sent by the reader. Saying “hello, I am here” takes less energy than changing nonvolatile memory. At the edge of coverage the hello may still work while the write fails. Move closer, improve the antenna path or increase legal reader power—but measure the result instead of assuming the read distance is the write distance.
There are two radio paths and several thresholds
The forward link carries RF energy and commands from reader to tag. The tag rectifies part of that energy into a supply voltage. The reverse link carries the tag's reply by modulating its antenna reflection—backscatter—so the reader can detect it.
An inventory succeeds only when the tag powers up, decodes the command and the reader receives the reply with enough signal-to-noise margin. A write adds another requirement: the tag must sustain the voltage and timing needed to alter nonvolatile memory and return a valid result. Depending on the system, either forward-link energy or reverse-link detectability can be the limiting factor.
The tag has enough harvested energy to start and participate.
The chip can access memory and form a reply while the reader can detect it.
The chip has enough energy margin to complete the programming operation.
The operation passes repeatedly across allowed samples, angles, surfaces and environmental variation.
A published sensitivity gap in real numbers
NXP specifies UCODE X at −26.2 dBm read sensitivity and −23 dBm write sensitivity. Because a less-negative dBm threshold represents more required received power, the write threshold is 3.2 dB higher.
Power ratio = 103.2/10 ≈ 2.09. In a simplified free-space forward-link model where received power falls with distance squared and every other condition is fixed, threshold range scales with 1/√power. The resulting idealized ratio is √(1/2.09) ≈ 0.69.
That does not mean every UCODE X system writes at exactly 69% of its read range. It is an engineering inference that isolates one published chip-sensitivity difference. Real installations include reverse-link limits, antenna patterns, multipath, reader behavior, regulatory limits and statistical reliability. It is even less appropriate to copy the ratio onto an undocumented changeable-TID IC.
Why distance alone is not a complete RF specification
| Variable | Mechanism | What to record |
|---|---|---|
| Conducted reader power | Sets transmitter output before cable/connector loss | dBm at the configured port, not “100%” |
| Antenna gain and pattern | Concentrates energy by direction and shapes coverage | Model, gain, polarization and tag position in the pattern |
| Cable/connector loss | Reduces power between reader and antenna in both practical setup and calibration | Cable type, length and measured/declared loss |
| Polarization | A mismatched tag and reader antenna couple less energy | Relative orientation and antenna type |
| Tag antenna match | Controls how efficiently the antenna delivers RF power to the IC | Exact inlay/card format and frequency region |
| Mounting material | Metal, liquid and nearby dielectric material can detune or absorb energy | Actual product surface and spacing |
| Multipath/interference | Reflections create peaks/nulls; other signals reduce receiver margin | Location, channel behavior and nearby structures |
| Operation | Inventory, Read and Write have different command/energy needs | Success rate for each operation, separately |
EIRP helps compare transmit setups, but does not finish the link budget
A simplified transmit-side quantity is effective isotropic radiated power: conducted reader power minus feed loss plus antenna gain, all in dB units. Regional rules may constrain radiated power and channel operation, so “increase power” is not an unlimited remedy.
EIRP (dBm) = reader output (dBm) − cable/connector loss (dB) + antenna gain (dBi)
The tag still sees position-dependent antenna gain, path loss, polarization loss and environmental effects. On the return path, the tag's backscatter and the reader receiver matter. EIRP is therefore a necessary comparison term, not a write-range prediction by itself.
A better test: power sweep before distance sweep
- 01
Fix the geometry
Mount one sample at a measured distance and orientation on the intended surface. Keep people and movable metal out of the test zone.
- 02
Find inventory and read thresholds
Step conducted power in small increments near failure. Run repeated trials and save the success count.
- 03
Find the write threshold
Alternate known patterns, verify raw readback, remove RF power and verify again. Count a write only after cold readback.
- 04
Add operating margin
Do not deploy exactly at the first-success threshold. Define margin based on expected orientation, product and environmental variation.
- 05
Repeat across distance and samples
Map the worst accepted case, not the most photogenic maximum.
How the special changeable-TID products should be specified
The supplied product information states a broad 0.1–10 m reader-dependent range and recommends a high-power reader. Generation 2 is reported to improve performance relative to Generation 1. Those statements are useful for selecting a sample; they are not a guaranteed write-distance specification.
A quotation-grade requirement sounds different: “At 30 dBm conducted power, with the named antenna and cable in the permitted region, 24 of 24 pilot samples shall complete 20 MB10 write/cold-read cycles at 0.5 m in the specified orientation on cardboard, with no adjacent-word change.” The numbers should come from the buyer's workflow, not from this example.
Where a shorter-write-range tag is a poor fit
Vehicle gates, toll lanes and passage portals
Distance, speed and uncontrolled orientation leave little room for a high programming threshold. Use qualified tags designed for the portal.
High-speed conveyor encoding
Short dwell time and a moving RF field can interrupt programming. Validate a production encoder and standard inlay first.
Low-power desktop or mobile readers
If the device cannot deliver the required field at the tag, software changes will not create energy margin.
Dense bulk writes
Population selection, collisions and uneven field strength complicate reliable per-tag programming.
On-metal or liquid products without a tuned construction
An ordinary thin inlay can detune severely. Use a tag designed and qualified for the material.
Any deployment specified only by “maximum meters”
Without operation, equipment, geometry and success rate, the number cannot be reproduced.
Read range answers “can I hear the tag?” Write range answers “can I deliver enough clean energy, for long enough, to change memory and verify it?” Treat them as separate specifications. For a special changeable-TID tag, the reliable write result on your hardware is the number that matters.
Before deploying: Review the applications where changeable-TID RFID is the wrong choice ↗
Primary reference and related testing
- NXP: UCODE X official product specifications ↗ — published −26.2 dBm read and −23 dBm write sensitivity values.
- NXP: UCODE 9 data sheet ↗ — example of a documented production UHF IC and memory implementation.
- Changeable-TID validation protocol ↗ — reproducible memory and RF acceptance procedure.