Subway Card: Chip, Protocol and Issuing Choices

time :2026-11-20 author :CBT scanning :

Subway Card: Chip, Protocol and Issuing Choices

A subway card looks like a plain piece of plastic, but a transit project succeeds or fails on four decisions: the protocol the gate already accepts, the chip family the fare system can manage, how cards are personalized and blocked, and whether the card body survives years of daily tapping through turnstiles. This guide follows the order those decisions actually happen, from a single metro line to a multi-operator city network with buses, gates and top-up machines.

Contents

1. What a subway card is, in fare-system terms

In an automatic fare collection (AFC) system, the subway card is the passenger-side credential. It holds a balance or a ride entitlement, is read at an entry gate and again at an exit gate, and is written back when the fare is deducted. That round trip — read, authorize, decrement, log — is what separates a fare card from a simple access badge that only returns a fixed number.

Physically it is an ISO/IEC 7810 ID-1 card, 85.60 × 53.98 mm and typically 0.76 mm thick, with an etched copper antenna coil and a chip module embedded between PVC or PET layers.

Two commercial models sit behind the same plastic. Closed-loop cards are issued by the operator and can only be spent inside that system. Open-loop cards carry a payment application and are accepted wherever the scheme is honoured. The choice changes chip family, security architecture and certification workload, so it is settled before any artwork is drawn.

2. Protocol first: ISO/IEC 14443 at 13.56 MHz

Transit gates worldwide are built around ISO/IEC 14443 proximity cards at 13.56 MHz, in either Type A or Type B signalling. Type A dominates legacy metro fleets and a large share of MIFARE-based deployments; Type B is common where a payment or government application sits on the same card. A gate reader is normally fixed to one type, so the card must match what is already installed rather than what is newest.

Older systems still run 125 kHz LF cards: shorter read range, one fixed identifier, low cost, and unable to hold a balance or run a cryptographic transaction on its own. Where a project extends an LF fleet, the practical route is a dual-frequency card carrying both an LF and an HF chip, not a wholesale reader replacement.

3. Chip selection: Classic, DESFire, CPU card or NTAG

Four families cover the large majority of transit tenders, and the difference is mainly how much cryptography the fare application needs.

  • MIFARE Classic — 1 KB or 4 KB, sector-based, widely installed. It uses a proprietary Crypto-1 authentication that has been publicly analysed, so it suits low-value, closed environments where the operator accepts that risk.
  • MIFARE DESFire EV2 / EV3 — AES or 3DES authentication, file-level access rights, transaction MAC and a flexible memory layout. This is the usual choice for stored-value cards that must resist cloning.
  • CPU card — a full microcontroller running a payment or transit applet, typically certified to a security level defined by the scheme. It carries the highest unit cost and the longest certification path.
  • NTAG — small memory, simple data model, low unit cost. Suitable for single-journey tickets and tourist passes where value is written once and read at exit.

For a side-by-side on the two MIFARE branches, see our breakdown of MIFARE Classic versus DESFire before the tender document fixes the chip.

4. Single-journey ticket, stored-value card and token

Three form factors serve different passenger flows, and mixing them is normal in a mature network.

A single-journey ticket is a thin paper or PET card, often 0.50 mm or less, sold for one ride and discarded at the exit; unit cost is low and volume high, so the print and encode line becomes the constraint rather than the chip. A stored-value card is a durable PVC or PET card reloaded at machines or by phone and held for years, so it needs key management and a body that resists scratching and bending. A token is a rigid disc read by a dedicated gate aperture — robust, unprinted, but tied to its own reader hardware.

Many tenders ask for one or two of the three. The mistake worth avoiding is specifying a durable card where the operator actually wants a recyclable single-ride ticket, which adds cost to every journey.

5. What the fare-system supplier needs before sampling

Card factories cannot choose the chip for a transit operator; the AFC integrator can, because the gate firmware and the security module define what is possible. Before a sample is cut, six items settle nearly every compatibility question:

  • Protocol and card type accepted by the gate (ISO/IEC 14443-A or -B, and any legacy LF layer).
  • Chip family and memory size already provisioned in the fare system.
  • Reader or gate model, plus its antenna position and rated read distance.
  • Whether the card must carry a second application, such as payment or a city pass.
  • Personalization format the system expects: sector layout, file structure, key set.
  • Expected card life and annual volume, which drive body material and MOQ.

Where the gate hardware is still open, our comparison of smart card readers and RFID proximity readers explains which reader parameters matter at the turnstile.

6. Security: SAM, mutual authentication and key diversification

Fare value is money, so the card is only as trustworthy as the keys behind it. Three mechanisms carry the load.

Mutual authentication. The card and the gate prove to each other that they hold the same key before a balance is read.

Key diversification. Each card derives its own key from a master key and its unique ID, so compromising one card does not expose the fleet.

SAM and transaction MAC. A secure access module inside the gate or the top-up machine holds the master keys and signs the transaction. The card returns a MAC over the new balance, so a replayed or edited message is rejected at the next gate.

Honest limitation: no contactless card is copy-proof. A determined attacker with physical access and time can often reproduce a card's data. The realistic goal is to make copying uneconomic — per-card keys, a signed transaction record and a blacklist that the gate checks at entry.

7. Card body and finish for daily turnstile use

A subway card is tapped, dropped and pocketed several times a day for years, and the body fails before the chip does in a large share of field returns.

Standard PVC laminated at 0.76 mm is the baseline: it holds a printed face, tolerates normal flexing and keeps the module flat. PET and PET-G are used where the card must resist heat or chemical cleaning, and they behave differently under lamination, so the antenna design is adjusted rather than copied. For premium or commemorative editions, our RFID PVC card guide covers overlay options, lamination temperature windows and the material grades that hold print registration.

Two body details get overlooked. First, module flatness: a module sitting a few tenths of a millimetre proud of the surface catches on gate slots and delaminates over time. Second, the antenna's position relative to the module — a coil that sits too close to the module edge loses coupling and shows up as intermittent reads on some gates but not others.

8. Printing and visual features that discourage copying

Printing on a fare card does two jobs: carry the operator's identity, and make a home-made copy obvious to staff at a glance.

Offset litho suits long runs and photographic artwork; silkscreen suits spot colours and heavy ink coverage on small batches; thermal transfer is used where variable data — card number, issue date, batch code — is printed at personalization time rather than at the card plant. Digital print sits between the two, economic for several hundred to a few thousand pieces with versioned artwork.

Visual security usually means a combination of guilloche patterning, microtext, a hologram patch or a colour-shift ink, plus a serial number in a font that is awkward to reproduce. On a transit card, the serial number matters more than the hologram, because it gives staff a reference point when a gate rejects a card.

9. Encoding, serialization and batch personalization

At production scale, the card is not finished when it is printed. Personalization writes the chip, prints the variable data and verifies the result, and it is the step that decides whether a batch is usable.

A typical transit batch carries a unique card number, a derived key, an initial balance or ride count, and a batch identifier. The encoding station reads back every field after writing and flags rejects rather than shipping them.

If the deliverable is a batch of blank cards that the operator will encode in-house, the factory still needs the chip's configuration file, so the blanks arrive with a known init state. Our note on RFID card numbers explains how UID, printed serial and encoded serial relate, which is where a large share of mismatch complaints start.

10. MOQ, lead time and what drives cost

Cost per card is set by five inputs, in roughly this order of impact: chip family, card body and thickness, print method and colours, personalization depth, and packaging.

Blank cards with a standard HF chip start from a few thousand pieces; printed DESFire cards with per-card keys start higher, because key generation and read-back verification take time per card. Single-journey paper tickets sit at the other end: low unit cost, high volume, with the print and encode line as the constraint.

Lead time splits into three blocks: chip procurement, which depends on allocation rather than on the card plant; production, which is days once material is in hand; and personalization plus verification, which scales with volume. Standard stock items can sample within 24 hours; a custom chip plus custom print plus per-card keys is a project, not a stock order, and should be planned with the tender schedule.

11. Standards and documents to check before ordering

Four documents answer the questions a procurement team usually raises, and they should be read before the specification is frozen.

  • ISO/IEC 7810 and 7816 for card dimensions and contact-chip behaviour, where a dual-interface card is involved.
  • ISO/IEC 14443 parts 1–4 for the contactless layer, including the anticollision and transaction sequence the gate relies on.
  • The chip vendor's data sheet for memory layout, authentication modes and endurance figures.
  • The AFC supplier's personalization specification, which defines sector or file layout and the key hierarchy.

Two habits save time later: ask for the endurance figure in the chip data sheet rather than a sales deck, and for antenna resonance measured on the finished card rather than the inlay. Both are testable, and both separate a compliant bid from a low-cost one.

12. A four-phase rollout for a new line

Phase 1, compatibility. Take the gate model, protocol, chip family and personalization format from the AFC supplier, and test three chip options on the real gate rather than a bench reader.

Phase 2, pilot batch. Several hundred cards, fully printed and encoded, run through gates at peak hours. Watch for intermittent reads, module wear and print scuffing, and record the gate that fails rather than only the count.

Phase 3, volume order. Freeze artwork, key hierarchy and packaging, then place the volume order with the personalization file locked. Any later change to the key set invalidates every card already written.

Phase 4, replenishment. Keep a rolling stock of blanks and a spare encoding line, because lost and damaged cards are a steady drain and replacement cards must match the live key set, not a new one. The card programme is a subscription, not a one-off purchase.

Frequently asked questions

Are subway cards RFID or NFC?

Both descriptions fit. A transit card is an RFID device working at 13.56 MHz, and it uses the same ISO/IEC 14443 air interface that phones use for NFC. The difference is usually the application on the chip, not the radio layer.

What frequency do subway cards use?

Most run at 13.56 MHz with ISO/IEC 14443 signalling. Some older bus and metro fleets still use 125 kHz LF cards, and those projects usually order a dual-frequency card rather than rebuilding the readers.

Can a subway card be copied?

With physical access and time, card data can often be reproduced; no contactless card is copy-proof. Per-card keys, a transaction MAC and a gate-side blacklist make copying uneconomic, which is the practical target.

What is the difference between a single-journey ticket and a stored-value card?

A single-journey ticket is a thin, low-cost medium used once and discarded. A stored-value card is a durable card holding a reloadable balance, with a chip that supports key management and a body built for years of use.

Can one card serve both the metro and the bus?

Yes, where both systems share a protocol, chip family and key hierarchy. Multi-operator schemes normally require a common specification and a central clearing arrangement before cards are ordered.

What thickness should a subway card be?

0.76 mm is the ID-1 standard for durable fare cards and matches common gate slot dimensions. Single-journey paper tickets are often 0.40–0.50 mm, and the gate must be specified for that thinner medium.

Can printing and encoding be done in the same order?

Yes, and transit orders often combine them. The factory needs the personalization file, the key set and the numbering range, and every card is read back and verified after writing.

What is a typical MOQ for custom subway cards?

Blank cards with a standard chip generally start at a few thousand pieces. Printed cards with DESFire chips and per-card keys start higher, because key generation and verification add time to every unit.

How long does a production batch take?

Chip allocation is the variable that usually sets the date. Production itself is days once material is in hand; personalization and verification scale with volume, so large batches should be scheduled against the tender timeline.

Do cards hold up in hot and humid cities?

PVC performs well in normal service; PET and PET-G are used where cards see heat or cleaning chemicals. Whichever grade is chosen, the antenna design is re-tuned for that material rather than reused unchanged.

Planning a fare card programme?

Send the gate model, protocol, chip family, personalization format, annual volume and destination country to info@chipbond.com. We reply within one working day with an engineering answer — antenna, chip configuration and body material checked against your AFC specification — rather than a price-only sheet.

CBT has produced RFID and NFC cards since 1996, with 20,000 m² of plant, 1.5 million cards per day, 99.8% yield and 24-hour sampling on standard stock, and runs twin bases in China and Thailand under one IQC/QA procedure. Peak-season volume can split across the two sites, the Thailand base offers an alternative origin where Chinese-origin goods face extra duty, and goods ship from whichever base sits closer to your port. Tell us the destination country and we will quote from the base that gives the lower landed cost.

Popular
Posts
Recommend Products