rfid reader selection method, frequency-band matching and analysis of deployment requirement elements
RFID Knowledge

rfid reader selection method, frequency-band matching and analysis of deployment requirement elements

PublishedShanghai Zejie Smart Technology Co., Ltd.

Point 0: The first thing in the selection is a match-up band.

The first step in the selection is to pair the frequency band with the scene, and then talk about power, antennas, interfaces. The same UHF reader, which is attached to the warehouse line, can read as much as 8 meters, and move to a metal-intensive machine processing workshop that can be within 2 metres. The frequency selection is right, it solves about half of the problem on the ground.

Element 1: Set the scene and then list the candidate model

You've often met a client with a model "Is the distance enough" and asked the opposite. Distance is the result, not the starting point. The real starting point is three variables: what material is labeled on, where the reader is placed, how many labels you want to read at a time. These three variables determine frequency, which determines the number of readers.

In early project assessments, Shanghai Zejie usually reviews these three variables with the customer, then shortlists three to five candidate models. This avoids recommending expensive equipment when a low-frequency application would meet the need.

Element 2: LF, HF, URF three frequency bands

The following is a comparison of common frequency bands with typical scenes, which can be directly referenced in the following scenario:

Frequency bands Frequency range Typical Read-outs Typical scene
LF Low Frequency 125–134 kHz less than 10 cm Animal trace, a cartoon, fuel.
HF HF High frequency 13.56 MHz 10 cm–1 m Library, ticketing, NFC payments, archives management
UHF UHF 860–960 MHz 1–10 m+ Storage, logistics, retailing, vehicle identification

The domestic UHF frequency band is allocated by the Ministry of Industry and Communications to 920-925 MHz, with a launch power ceiling generally not exceeding 33 dBm (joined 2W), which is a condition for compliance and cannot be crossed when programming.

Element 3: Power and antennas, match more than stack parameters

Many of the selected failures are not problems with the device itself, but they are overfilled. For example, in machine workshops, you can move the UHF reader output to 33 dBm, and then add 12 dBi to a high-benefit orientation antenna, which is used to absorb most signals, and the reading distance shrinks.

Several experience values that can be used directly:

Element 4: Interface protocol leaves 1 to 2 years of extended space

Readers do not work in isolation, they report data to the upper level, pick up PLCs and feed information to the top. The interface is incomplete, and later either a gateway or equipment replacement costs much more than the start time cost of several hundred pieces of pairing.

Common interfaces divided by level:

For projects with more than four readers, choose Ethernet-equipped models and allow for Power over Ethernet (PoE). This alone can halve cabling and power installation labor.

Point 5: Multiple labels, with 30% to 40% excess

The client sees the chip manual with "200 labels per second" and thinks it's enough. The fact is, under a dense scenario (the whole tray 200 SKUs also enter the reading area) the label will actually swallow up to 50%-70% of the nominal value because of its orientation, stacking cover, signal interference.

Rule of thumb: usable throughput ≈ rated speed × 0.6. For peak traffic estimates, also apply a factor of 1.3 to allow headroom. Both factors have been tested on several production lines and have proven fairly reliable.

Point 6: Field testing is a mandatory link, not just laboratory data

All the parameters are right, and you're going to try the reading on the real scene. Focus on three things:

  1. Is the read rate at least 99% when tags are attached to actual materials? This is the minimum most production lines will accept.
  2. The true rhythm of the current line (e.g. 1.2 seconds/spm) has the drop rate skyrocketed?
  3. Runs continuously for more than 8 hours, observing temperature drift and leakage trends

Test sample tags at different orientations. In production, workers won't always apply tags exactly to specification, so cover front-facing, side-facing, and angled positions. Dropping from 99% to 95% sounds like only four percentage points, but it means missing 400 items in every 10,000. A lab test is unlikely to catch that.

Point 7: Frequency planning is not sufficient for multiple readers ' networks

With three or more UHF readers in one work area, interference is a common source of failures. UHF has multiple channels, but available frequencies in China are limited (840–845 MHz and 920–925 MHz, with N channels at 200 kHz each). The text recommends offsetting neighboring readers by at least 5 MHz and gives 920.625, 925.125, and 924.375 MHz as examples of frequency hopping. Combined with carrier sensing and time-division multiplexing, this can reduce interference to acceptable levels.

If you are not helped to plan your frequency, you will probably have to return to work. This step will typically be a frequency distribution map that will indicate the channel, power and antenna direction of each device, so as not to make any direct calls from the field.

Conclusion: a list of options that can be used directly

The sequence of the seven points above is a selection process that follows:

  1. Clear parameters for the scene: label material, ambient temperature, reading quantity
  2. The fixed frequency segment: LF / HF / UHF Three-Selection One
  3. Estimated distance: Combined antenna gain, power, real decay, 1.5-2 times more
  4. Select interface: On-site bus + network layer, expansive priority
  5. Calculate throughput: use 60% of the rated figure as the realistic rate, then allow 30% headroom for peaks.
  6. Test on site with actual materials at actual production speed for eight continuous hours.
  7. Group planning: Stagger frequency + Time Schedule + Output frequency table

After this 7-step purchase order, the first line of projects can avoid more than 80% of common pits. The remaining 20%, mostly on wiring and power lines, that part has nothing to do with frequency, but it's worth looking at in advance.