HF vs. UHF RFID Readers: How to Choose and Test Each Option
The selection first asks: What exactly is this identification to prove?
A plant may need two types of RFID at the same time: a workspace that identifies this material box in front of it, and the exit door checks the whole cargo. The former wants the next-door material box not to be selected, the latter wants the vehicle to read as much as possible when it passes. The choice between high-frequency readers and ultra-high-frequency readers depends first on where one of the items is going. This is a more common source-free HFC and UHF program, discussing frequency bands, group reading, material impact and project costs, and giving a decision-making method that can be validated by actual samples.
Frequency and coupling: Why is it far from appropriate to read?
Frequency influences how energy and information are delivered, but does not determine the effective reading distance of a system in isolation. Readers, label chips, antenna sizes, installation materials, direction and field setting need to be discussed. Define "HF can only read on top, UHF can read only remotely" and miss practical options such as HF multi-label systems and UHF per diems. The judgement should be placed in a specific product combination, not a frequency segment name.
The standard HF working frequency is 13.56 MHz, and the typical passive program uses a magnetic field coupling between the coil. ISO/IEC 14443 and ISO/IEC 15693 are a system of protocols that need to be distinguished, and the reader who supports the former does not necessarily support the latter. Close range cards, controlled areas are identified and multi-labelled, can be designed using different antennas and product points. Short distance is sometimes an advantage: operators place an object in a workplace, and systems only need to confirm it reliably, and do not want to see the item next to be considered completed.
The common UHF logistics label uses a reverse-dispersible communication, with long-range area identification and a specially designed near-site application. The EPC UHF Gen2 is the usual protocol check for ISO/IEC 18000-63. No matter how far you want to read it, you can ignore the possibility that an adjacent object will be collected. The project needs a reading area appropriate to the operational boundary: the target identified is identified, and the person not supposed to be involved in this task does not trigger business actions by being read by accident.
Read how to compare: Number, borders and tasks take time
When comparing the prototype, you should first write a list of the people to read, labels, object mode and the business actions to be performed. Otherwise, one device returns repeatedly with a large number, which appears fast, but may also count the next shelf; the other only returns a small number, and completes all the check on the assigned workspace. Readable numbers are meaningful only within the mission boundary.
Group reading: HF can also identify multiple labels, but the terms of the mission are clear
The UHF is often used for large-scale mass identification, but the HFC does not mean that it can only be read by one. ISO/IEC 15693 also supports collision prevention, and “multilabel” relies on protocol organizations to respond, not all of them speaking at the same moment. The actual effect still depends on label spacing, stacking, coupling, data length and allowed time. If the goal is to read dozens of books, this group should be tested directly; it cannot be calculated successfully with a single card or pre-rejected with the word “high frequency”. It is also necessary to distinguish between multiple identifiers and full reading of each label's assigned data; both do not use the same amount of input commitment.
Stability: Full reading in the area of identification, less reading outside the area of identification
The effects of metals and liquids are analysed separately. The vortex loss and the loss of antennas that may occur when HFCs approach a metal cannot be considered as a means of identifying penetrating metals; the UHF can also be affected by metallic reflection, label mismatches, liquid absorption and peri-pee. Specially designed anti-metallic structures, quarantine material or labelling positions can improve part of the conditions, but must be verified. Liquid loads, containers and label locations are equally important, and “high frequency is not afraid of liquids” are not sufficient to form the basis for selection.
Speed: Read numbering, reading and writing are not the same tests
Read a logo, read a user data, complete authentication and write back, and spend different time. If the project needs to write batches and process states, you cannot use a read-only number demonstration rate as acceptance commitment. It is recommended that the "mission trigger-collect-data validation-business confirmation" be used, and that read-and-read, retry and manual processing be included in the results. For example, 98 readable labels on the first round, not equal to 100 items have been removed from the library; whether the remaining 2 are to be opened to determine how the whole batch will take place. The figures here are only for test calibres, not for any equipment.
How to account for costs: the cost of comparing a correct operation
The cost comparison should be made using the same object, the same number and the same delivery boundary. Only the reader's offer does not answer whether the project is economical: some programmes are cheap in terms of equipment, but they require re-adjustment of work spaces; others have higher unit price, but can be used continuously during actual cleaning or swing cycle. First, it is clear whether a one-time screech or a circular load, then the input and long-term expenditure.
Beyond the chip and antenna, tag costs include packaging, printing and encoding, application checks, heat or wash resistance, damage replacement, and unlinking at retirement. Finished adhesive tags and inlays are not quoted on the same basis. An inlay usually consists of the chip, antenna, and substrate and can be converted into different finished formats. Compare HF and UHF tags with matching materials, dimensions, order quantities, and test requirements. Comparing ordinary paper tag prices with industrial housings does not yield useful results.
The system cost should also include antennas, cables and installation, area control, interface development, debugging, transport peacekeeping anomalies. It can be calculated over the same period: equipment is shared with implementation during each correct business operation = time frame, consumables, maintenance and manual processing costs and the number of operations that are correctly performed. If a programme is collected faster but requires extensive reconciliations every day, these hours will be added. The saved operational time will only be converted to reduced overtime, released capacity or verifiable error losses, so it is appropriate to account for the corresponding gains, not to convert all the savings into cash.
There are different options in the same industry: books, tools and trays, for example.
Industry names help to find a candidate, but they cannot replace mission analysis. Library loans confirm that the same action is not taken as export-proofing, and tool identification and tool cabinet bulk count are not. They should be reduced to two categories of target, then tested for specific objects, so that they do not follow "a frequency segment used in a certain industry" directly.
- Controlled appearances interact with individual piecesThe door-ban card, membership certificate, touch service, job identification may be assessed from the HF candidate ' s program; need for certification, multilabelling or anti-misselection is determined by operational risk and safety cannot be inferred from short reading distance.
- Area collection and batch interface(b) The cargo container access warehouse, clothing inventory, logistics transfer and flow equipment count can be assessed from the UHF candidate programme; missing reading, serial reading, direction and mission relevance are verified at the same time and no numbering in the area is directly considered a stock change.
For example, the box is handed over to the identified operator for confirmation of the intended subject, and the whole box is re-examined to the full object in the identification box list; both need to prevent a string reading and bind error, not just by distance. Read the tray: Read the tray label only to show that the tray is marked and that each of the contents is complete. If you use the tray to be bound, untrace the tray, exchange the goods, and remove the envelope, it must be maintained; Otherwise, both the HF and the UHF are just spreading a false load list quickly.
Four easy to misjudge.
Many of the errors in the RFID selection are due to incomplete testing conditions, rather than inadequate equipment performance. The following four problems, which should be recorded and calibrated at the prototype stage, are more useful than continuing to show normal success: a reverse example can indicate when the program needs to stop and hand over manual confirmation.
Only the nominal reading distance, no comparison of field work
The maximum distance on the specifications is usually for a specific label, antenna, power and plaque. The field must cover full load and empty load, overpack changes, different entry directions, work at the adjacent place at the same time, and the maximum allowed speed of traffic. It is recommended that separate records be recorded for the first round of label coverage, the number of off-site miscollections, the number of times business errors are accepted and manual interventions take time. Assuming 100 copies should be read, 98 of them should be read, and 5 in the neighbourhood; "Return 103 different numbers" does not indicate completion rates exceeding 100 per cent. This group assumes that the data reveals wrong statistical calibration, which does not represent the result on-site.
Focus on the URF polarization, ignoring the direction of the HFC loop.
UHF linear polarization, the Circumpolarization and the direction of label antennas affect alignment; Circumpolarization usually helps to deal with some of the changes in direction, but does not eliminate the problems posed by labels, or by moving towards and shielding. The HF is also affected by the relative direction of the circle, by the shift in position and the conditions of convergence, which may be difficult to read when a label is up or away from a valid area. It should rotate the actual label, change the relative location and keep the worst case, rather than just measure the most advantageous one. ST AN 2972 applies notes to check basic problems of NFC antenna alignment and environmental harmonization.
Use the NFC as a universal key to all HF tags.
A phone's built-in NFC cannot directly count UHF tags, but operation at 13.56 MHz does not mean the phone can read any data on a tag. Check the phone OS, supported tag technology, NDEF format, app permissions, and authentication requirements. Reading a tag identifier is different from opening a web page, reading business records, or changing protected content. For a tap-to-open service page, test the actual target phones with the final data format instead of relying on a single success on a developer's phone.
Equivalent protocol as data availability
The frequency and air interface protocol match is a starting point, and checks the specific chip command, memory structure, access rights, key management and up-to-top interface. Some devices can count to label, but do not have an open command to complete target authentication or writing. For long-term systems, assign coding, export format, replace tags with a device replacement. It is easy to move depending on the full data and interface design, not just how many standards are written in the product information.
Let the sample test give a retroactive selection basis
The same representative sample can eventually be prepared for the two candidate sets, which will record normal operations with the same task scripts, material changes, neighbourhood interference and network restoration. First set unacceptable errors, such as miscarding of objects, and release them automatically; then compare the time of mission after meeting these constraints, manual retreat and long-term cost. If the two frequency segments are suitable for different nodes, a combination of HF sites may be used to verify the interfaced collection with the UHF, but the identity and operational status of the items should be harmonized. The selection report should explain why this option is being chosen under current conditions, and which conditions need to be re-measured after they change, not just leave a model of equipment.

Ms. Lei
Mr. Pang