RFID technology guide for buyers, engineers and sourcing teams

A passive RFID chip harvests power from the reader's electromagnetic field via its antenna coil or dipole — a process called RF-to-DC rectification. When the reader transmits radio energy at the target frequency (125 kHz, 13.56 MHz, or 868-915 MHz), the tag's antenna resonates and induces alternating current, which the chip's front-end rectifier converts to a small DC voltage (typically 1-3 V) sufficient to run the digital logic and respond to the reader. This is why passive tags have effectively unlimited shelf life (10-20+ years) — there's no battery to deplete — but read range is bounded by reader power and antenna geometry rather than tag capability alone. Active and semi-passive tags use onboard batteries for extended range (10-100+ meters) but trade battery lifespan (3-5 years) for the capability.

UHF Gen2v2 (ISO/IEC 18000-63) uses a probabilistic slot-based anti-collision protocol called Slotted Aleph-Q. The reader broadcasts a "query" with a Q parameter (2^Q slots). Each tag in the field picks a random slot. If one tag answers in a slot, the reader reads it. If multiple tags answer in the same slot, the reader detects a collision and issues a new query with higher Q (more slots, fewer collisions). The process repeats until all tags are inventoried. A modern fixed reader (Impinj Speedway R700, Zebra FX9600) can inventory 700-1,200 tags per second in a dense field. For HF (ISO/IEC 14443), anti-collision is tree-based with slightly different logic but similar throughput at much shorter range. For LF (ISO 11784), anti-collision is limited — most LF readers are designed for single-tag presentation, which is why LF is not used for bulk-read inventory.

The TID is a factory-programmed unique identifier burned into the chip at manufacture — it cannot be rewritten or cloned, unlike the EPC memory bank which is writeable. For security-sensitive deployments (high-value asset tracking, anti-counterfeit, brand authentication), the TID is the tamper-proof anchor that ties a specific physical chip to a specific record in the customer's database. Even if a counterfeit tag is written with the same EPC, the TID will not match. Procurement specs should require TID read-back on sample verification for any program where cloning is a business risk — including luxury goods (Digital Product Passport), pharma, high-value tool crib management, and regulated access control. Not every chip exposes an unchangeable TID; NXP UCODE 9 and Impinj M730/M800 both do, while some older chips do not.

Yes, because UHF Gen2 chips are broadband (typically 860-960 MHz internally) and respond to whichever sub-band the local reader transmits on. The limiting factor is the antenna — antennas designed for the FCC 902-928 MHz band have peak efficiency there but still work at 60-80% efficiency in the ETSI 865-868 MHz band, and vice versa. For global supply chain deployments (export cartons, shipping containers, RTPs moving across regions), RFIDAK builds "worldwide" tags with broadband antennas optimized for 860-960 MHz rather than region-peaked. Trade-off: 10-15% lower peak read range per region vs a region-specific antenna, acceptable for most logistics applications. For region-locked deployments (only North America, only EU), region-specific antennas deliver maximum range. Specify "global read" vs "region-specific" in your RFP.

A modern UHF sticker (Impinj M730 inlay, AZ-9640 antenna or equivalent, 70×17 mm) on a dry cardboard surface typically reads at 6-10 meters with a 2-8 W EIRP fixed reader and a directional antenna. Variables that degrade range: wet cardboard (-30-40%), metallized ink in the printing (-50-70%), proximity to other RF-absorbing contents (liquids, foil, electronics), tag misorientation relative to the reader polarization, and multi-path interference in crowded warehouses. Real-world deployment range for dense-read applications (warehouse gate, dock door) is typically engineered to 3-5 meters as the conservative working range — not the theoretical maximum — to ensure 99.5%+ read rate even under worst-case conditions. Always pilot-test with your actual product-packed boxes in the actual receiving dock before finalizing the reader layout.

Physically, yes — anyone with a standard reader at the right frequency can detect the presence of an RFID tag and read its EPC / UID. For high-risk applications, four mitigations apply. (1) Use chips with password-protected read (MIFARE DESFire, Plus EV2, NTAG424 DNA) so the EPC is encrypted. (2) Use chips with kill commands (EPC Gen2 supports a "kill" command that permanently disables the tag after the point of sale — standard for retail apparel Walmart mandate). (3) Use Faraday-cage packaging for high-value shipments. (4) Store only a random ID on the tag — map it to customer data on your secure server — so even if the tag is read, the attacker gains no useful information. EU GDPR applies to RFID data where it can identify an individual (e.g., RFID-enabled ID cards); consult legal counsel when deploying personal RFID in EU markets.

In 2008, researchers at Radboud University Nijmegen (Garcia et al., USENIX Security 2008) published a full cryptographic break of Crypto-1, the 48-bit proprietary stream cipher used in MIFARE Classic. The attack allows an adversary with a Proxmark or similar hardware ($200-800) to clone a MIFARE Classic card in minutes by reading it, cracking the key, and writing a new identical card. Practical impact: hotel locks, gym access, legacy transit fare cards using Classic can be duplicated at scale. Not every deployment has migrated — Classic still represents 20-30% of installed HF access control worldwide in 2024 — because migration requires replacing both cards AND readers (or moving to MIFARE Plus EV2 which emulates Classic but adds AES-128). For new deployments: specify MIFARE DESFire EV3 or NTAG424 DNA with AES-128. For legacy deployments: plan migration within 3-5 years as insurance carriers increasingly require AES-grade security on physical access.

Depends heavily on environment. Indoor-only stationary deployments (warehouse rack, office equipment, server rack): 15-25 years — limited by chip memory endurance (100,000 read/write cycles for most NTAG, 500,000+ for DESFire) rather than physical degradation. Outdoor exposed UHF tags: 5-12 years, limited by UV, thermal cycling, and adhesive degradation (not the chip). Industrial laundry textile tags: 200-400 wash cycles at 85-90°C = 2-4 years in daily hospital rotation. Autoclave-sterilized PPS tags on surgical instruments: 500+ autoclave cycles, 3-5 year functional life. On-tire UHF tags: 80,000-150,000 miles, 3-7 year fleet life. For ROI modeling, amortize tag replacement into the per-asset cost; the chip usually outlives the asset it's attached to, not the other way around.

No. iPhone (and Android phones) integrate NFC modems that work at HF 13.56 MHz only — they cannot transmit or receive at UHF 860-960 MHz. UHF needs a dedicated reader. iPhones CAN read NFC tags (NTAG family, MIFARE DESFire EV3, NTAG424 DNA) via the built-in Core NFC API (iOS 13+). For UHF in a consumer-facing application, options are: (a) supply customers with a dedicated UHF reader — not realistic for B2C; (b) use Bluetooth sled UHF readers that connect to phones (for field workers, not consumers); (c) switch to NFC for the consumer-facing touchpoint and use UHF separately in warehouse workflow. For retail programs wanting both "consumer tap at store" AND "warehouse bulk read" on one tag, dual-technology hybrid chips (RFID + NFC on one antenna) exist but add 40-80% cost per unit.

UID (Unique Identifier) is a general term for any number that uniquely identifies a tag. EPC (Electronic Product Code) is specifically a GS1-defined encoding standard used on UHF tags for supply chain, built around a 96-bit identifier structured as Header + Company Prefix + Item Reference + Serial Number. The EPC is writeable — you program your company's EPC onto the tag during personalization. HF tags use different terms: MIFARE Classic has a 4-byte NUID (Non-Unique Identifier, since collisions are possible), MIFARE Plus / DESFire / NTAG have 7-byte UIDs (genuinely unique across all chips ever made). For programs needing anti-counterfeit, also specify the TID (see Q3) which is factory-burned and unchangeable. Rule of thumb: in a UHF-RFID procurement spec, write "EPC". In an HF / NFC spec, write "UID" or "7-byte UID". Getting this wrong makes you look like a novice to vendors.

For pacemakers and ICDs: commercial RFID readers are thoroughly tested under IEC 60601-1-2 and ANSI/AAMI PC69 standards; contemporary (post-2014) implantable devices are hardened against RFID interference at normal reader power levels (<= 4 W EIRP for UHF). Recommendation: direct reader contact at <20 cm with a pacemaker should still be avoided — hospital and warehouse reader installations routinely include signage. For aircraft: RFID in aviation is an FAA-regulated application. FAA AC 120-108A specifically permits passive UHF RFID in aircraft cabin (luggage tracking, catering cart identification, life vest inspection) with certified reader installations. Active RFID (battery-powered, higher TX power) requires additional certification. Pre-2000 cardiac pacemakers and older avionics can be susceptible to strong EMF fields — follow hospital / airline specific protocols when in doubt.

Rule of thumb: outsource the tags, hardware, and integration to specialists; own the data layer and application logic. RFID vendors like RFIDAK operate at the intersection of chip supply chain, antenna engineering, encapsulation, and QC — skills that take years to develop and do not scale economically below 50K+ tags/year per SKU. Hardware integration vendors (Zebra, Impinj systems integrators, Nordic ID partners) handle reader deployment and real-time middleware. What YOU should own: the database that ties tag EPC/UID to business entities (assets, garments, customers, inventory), the business rules (when does an event trigger an action), and the reporting layer. This "bring your own data layer, outsource the physical layer" architecture is standard for enterprise deployments at Walmart, Boeing, Maersk, major hospital systems, and global hotel chains. For pilots under 5K tags, RFIDAK can provide end-to-end solutions including middleware reference code.

No. RFID chips store their ID and user data in non-volatile semiconductor memory (EEPROM or FRAM), not on magnetic media — so neither household magnets nor industrial lifting magnets will erase or damage them. The confusion comes from magnetic stripe cards, which genuinely are wiped by strong magnets. What actually kills RFID tags in practice: physical damage to the antenna (creasing a label across the antenna trace), electrostatic discharge during handling, sustained temperatures beyond the chip's rating, chemical exposure that delaminates the inlay, and the intentional EPC Gen2 kill command. If tags arrive dead from a vendor, the cause is almost always converting damage or ESD in production — which is why the QC read-verification percentage belongs in your RFP (see section 12).

Effectively, yes. Metal reflects RF energy and detunes the tag antenna, so a tag wrapped in aluminum foil generally cannot be energized or read — the same shielding principle behind RFID-blocking wallets and Faraday bags. For buyers this cuts two ways. Defensively: foil-lined packaging, metallized labels or foil-stamped artwork on your product can silently kill read rates — always test printed production samples, not blank ones. Offensively: this is exactly why bare-metal assets need purpose-built on-metal tags, which use a tuned spacer or ground plane to turn the metal surface from an enemy into part of the antenna. If metal, liquids or foil packaging appear anywhere near your tag, that environment note belongs in the sample-test plan (see section 7).