RFID Tag Key Fob OEM Orders: The Spec Sheet That Decides Whether Your Batch Works

A sample arrives at your office. On the desktop reader it presents cleanly at around six centimeters, the logo is sharp, the color matches the Pantone reference. You approve it. Twenty thousand pieces ship, and three weeks later the installer calls from site: on the wall-mounted panels the same part needs to come within about two centimeters before anything happens, and roughly a fifth of the doors take two or three presentations. Nothing was wrong with the sample. Nothing was wrong with the production run either. What was wrong was the specification, and on a custom RFID tag key fob order the specification is a shorter document than most buyers think.

 

 

This is not an article about choosing between ABS and epoxy. It answers one question: what has to be locked in writing before the first mold shot, and what the factory will assume in place of each item you leave blank.

 

Why a Quote Request Without a Data Package Is Just a Price Guess

 

Most enquiries that reach a factory contain a quantity, a color, a frequency and a logo file. That is enough to produce a number. It is not enough to produce a product that works in your system, because the variables that decide whether the batch functions on site are not in that message at all.

 

Here is what happens to each blank field. A factory quoting a bulk RFID tag key fob order cannot leave a field empty and cannot stop production to ask about all of them, so every unanswered item is resolved the same way: with the configuration that runs on the line most often. Frequency unstated becomes the highest-volume stock. Chip unstated becomes whatever is bonded that week. UID length unstated becomes four bytes, because four-byte parts are cheaper and more common. Encoding unstated becomes sequential numbering from one. None of those defaults is wrong in isolation. All of them are wrong if your site already has credentials enrolled.

 

That is the whole argument of this article, and it is why an order that was priced correctly still gets rejected at handover. A complete data package covers chip identity, UID structure, encoding format and number range, housing and tooling ownership, environmental rating, and batch strategy for repeat deliveries. Each section below is one of those categories, expanded into what a factory does with the information and what it does without it. If you are still at the stage of choosing a product rather than specifying one, the custom RFID tag keychain buyer's guide covers selection first, and the process that turns a finished spec into a production run is set out on our OEM and ODM manufacturing page.

 

 

UID Byte Length and Encoding Format: Where Whole Batches Die

 

Two numbers decide whether a pre-encoded RFID tag key fob can be enrolled in your system, and buyers routinely omit both.

 

The first is UID byte length. A four-byte UID and a seven-byte UID are not two sizes of the same thing, and a data set captured from one cannot simply be written onto the other. The failure has a signature. Writing a mismatched block zero returns a checksum complaint, recorded verbatim in practitioner write-ups on phone-based cloning as BCC of block 0 is not valid, and the tag is typically bricked rather than merely rejected (timdows.com). One tag is a nuisance. The same operation applied across a delivery is scrap. When byte length is not stated on the order, it is discovered by your installer instead of by your QC.

 

Now the format itself. Under the SIA 26-bit Wiegand standard the facility code occupies eight bits and the card number sixteen, which caps the format at 255 usable facility codes and 65,535 card numbers within each one, and the card issuer rather than the format owner controls which of those combinations goes out (HID format white paper, PDF). That ceiling is the entire reason collisions happen. It is not theoretical. Generic bulk fobs on open marketplaces are routinely listed with the facility code fixed at a single published value: 127, 98 and 99 all appear as stock settings, with card numbers running a preset range from 00001. A buyer topping up an existing site with "the same kind we bought last time" can therefore enrol a credential that duplicates one already in the panel. No goods-in test catches it. It surfaces as a door that opens for the wrong person.

 

Which is why format, facility code, start number and block size are not optional fields on an RFID tag key fob quotation. They are the order. What the handover looks like on the factory side is unglamorous and worth stating plainly. The buyer sends the Wiegand format and bit length, the facility code, the starting number and the allocated range. For block-level encoding, a dump file or block map from the existing credential replaces those fields. The factory then encodes a first article, and the buyer reads that first article on their own reader and confirms before the line runs. In practice the first-article step costs a day or two of schedule; a number range discovered to be wrong after delivery costs the entire lead time again, plus the scrap. Personalization can also be split: the Tajikistan card programme Syntek supplies at roughly two million pieces a year takes delivery encoded and then debosses names and numbers in-country, which is a reasonable model whenever the buyer holds data it would rather not send abroad.

 

Housing, Tooling, and the Order That Doesn't Match the Last One

 

Everything above concerns what your reader sees. This section concerns what your customer sees, and it is the complaint that actually damages a distributor relationship: the second delivery doesn't look like the first.

 

Standard housings across the industry are near-identical at a glance and meaningfully different in hand. Wall thickness, keyring aperture, surface texture and color tolerance all shift when a batch is produced from a different mold, and end users notice. For a private label RFID tag key fob mold programme this is not cosmetic. It reads as a quality drop to the person carrying it, and it makes your product straightforward to counterfeit for anyone able to buy the same generic shell.

 

The defence is a dedicated mold, and the contractual term to check is ownership. Syntek's published position is that a mold developed for a customer belongs to that customer, cut to ±0.02mm, with printed color held to ΔE<1 across repeat runs. That combination is what makes shell geometry a brand asset rather than a shared commodity. There is a documented case behind it: a French supplier of building access badges has run the same programme for around five years at roughly half a million pieces annually, and the mold was modified to carry a logo unique to that customer specifically so its brand would be harder to counterfeit.

 

Material selection sits underneath this decision rather than above it, and the guidance is narrower than most catalogues suggest.

 

Requirement driving the choice	Housing	What you are actually buying
High-volume issuance, cost-led, dry indoor use	ABS injection moulded	Lowest unit cost, widest colour range, silk screen and laser marking both viable
Wet, humid or chemically exposed environments	Epoxy encapsulated	Sealed construction rated to IP67, tolerates poolside, spa and washdown areas
Brand-visible or premium issuance	PU leather over ABS core, metal, wood	Perceived value and tactile differentiation; unit cost rises substantially

 

If the shell is settled and only chip and encoding are open, the existing RFID key fob product range is the faster starting point for a custom RFID tag key fob order, and the flat RFID key tag format covers cases where thickness on the ring is the constraint.

 

Note what is not in that table: read range. Housing barely affects it. The thing that does is the subject of the next section, and it is where most warranty disputes actually originate.

 

Read Range Belongs to the System, Not to the Fob

 

Return to the batch that failed at the door. In almost every case of this kind the credential is fine and the physics of the installation are not.

 

Low frequency credentials in a small form factor need their antenna wound on a ferrite core to concentrate flux into a coil that would otherwise be too small to power the chip, and a winding on a ferrite core is directional: performance peaks when the core axis aligns with the reader's antenna axis and falls away as the angle opens (EDN). A fob lying flat on a bench, presented squarely, is being tested under best-case geometry. The same 125kHz fob dangling on a keyring at whatever angle the user's wrist produces is not.

 

 

The reader side is worse. Mounting an inductive reader against a steel surface loads the antenna and shifts its self-resonant frequency away from the drive frequency, and the loss is not marginal: patent literature on compensating for exactly this effect describes typical short-range readers dropping to a fraction of their nominal range on a steel door, recoverable only by introducing a ferrite shunt behind the unit (USPTO). Add the keys, coins and metal fittings sharing the ring, and a site can lose most of its margin without a single defective part involved.

 

The procedural fix costs nothing and belongs in your sampling protocol. Test the sample on the actual reader model, in the actual mounting position, with the fob on a realistic keyring rather than presented bare. Any sample approval that skips those three conditions has verified the part and told you nothing about the installation.

 

The procedural fix costs nothing and belongs in your sampling protocol. Test the sample on the actual reader model, in the actual mounting position, with the fob on a realistic keyring rather than presented bare. Any sample approval that skips those three conditions has verified the part and told you nothing about the installation.

 

That holds on one condition, though: the factory can only reproduce your site if it knows what your site is. Send the reader model and manufacturer, and a photo of the mounting surface and its surroundings, together with the sample request. A factory that has both can bench-test against the same reader family and flag a geometry problem before tooling rather than after installation. Where the standard coil genuinely cannot deliver the margin, coil geometry is itself a specifiable item, as in the card programme mentioned earlier where the coil and chip were both purpose-built for the application.

 

Duplicate and Clone Fobs: Drawing the Line Before You Order

 

There is a category of product every buyer in this market encounters: writable-UID stock, sold under names like Gen1A, CUID and FUID. These exist because the UID on a standard credential is fixed, which makes legitimate replacement difficult, and the market answered with tags whose manufacturer block can be rewritten.

 

The uses are real. A property manager reissuing a lost credential without recalling every fob in the building, an integrator migrating a site from one panel to another, a locksmith serving residential customers: all of these need a duplicate RFID key fob and none of them are doing anything improper. The line is drawn by who authorises the copy, not by the technology, and a supplier that pretends the category doesn't exist is simply selling it under another label.

 

What matters more for your specification is knowing when UID-based identification is adequate, because the answer is genuinely scenario-dependent rather than universally negative.

 

In a gym or leisure club where the credential opens a turnstile and identifies a member for billing, a UID-only 125kHz credential is a rational choice: the asset behind the door is low value, the membership database is the real control, and cost per issuance dominates. In a multi-tenant office building where the same credential grants access to floors leased by different companies, UID-only identification is inadequate, because anyone who can read a UID in a lift lobby can reproduce it, and there is no cryptographic step to fail. In a hotel or serviced apartment, the deciding factor is usually the lock platform already installed rather than a free choice, and the specification question becomes which protocol the existing locks accept before it becomes a question about the fob at all.

 

What a Factory Needs From You Before Tooling Starts

 

Everything above condenses into one document. Send this with your enquiry and the quotation that comes back is a real quotation rather than a placeholder.

 

Field	What to state	Consequence if omitted
Frequency	125kHz / 134.2kHz / 13.56MHz / UHF 860–960MHz	Supplier assumes its highest-volume stock
Chip reference	Exact part number, not family name; state if a specific reference must hold across repeat orders	Silicon may change between batches without notice
UID byte length	4-byte or 7-byte, and whether UID must be sequential	Enrolment failures and bricked parts
Encoding format	Wiegand format and bit length, facility code, start number, number range, or block-level data map	Collisions with credentials already enrolled
Housing and tooling	Standard shell or dedicated mold; confirm mold ownership in the contract	Second order differs visibly from the first
Marking process	Silk screen, laser engraving, laser serial number, QR or barcode, color references	Color drift and unreadable variable data
Environmental rating	Indoor, outdoor, wet, chemical, high temperature	Housing selected on price alone
Quantity and batch plan	Total volume, call-off schedule, number range allocated per batch	Number ranges overlap across deliveries
Sample and approval	Reader model and mounting position used for testing; who signs off	Approval on bench conditions only

 

Fill that table in, add your reader model, and send it before you ask for a price: our engineers return chip and encoding recommendations against it first, which is what a custom RFID tag key fob manufacturing review is supposed to produce. A quotation issued against a completed table is a commitment; one issued against four lines of email is an estimate.

 

Two rows are worth restating because they are the ones most often left blank: chip reference and encoding data. A factory can absorb a vague answer on packaging. It cannot absorb a vague answer on either of those without guessing, and its guess will be its standard configuration.

 

How to check that an RFID tag key fob factory can actually hold that spec

 

A specification is only worth what the supplier behind it can hold, so the second half of this is verification rather than drafting. Syntek was founded in 2006 and produces in a 3,600㎡ plant with five production lines, over 200 staff and daily output in the hundreds of thousands of units, running its own ABS and TPU molding, mold-making line, chip bonding, copper coil winding and silk screen printing in house. That is what makes a single stated chip reference possible to hold from order to order rather than reassembled from three sub-suppliers each time. Before shipment, every batch passes at least three full inspections against four stated criteria: frequency, read range, data integrity and appearance. The factory holds CE, ISO 9001:2015 and ICAR certification and carries a six-month warranty position. None of that has to be taken on trust: standard models ship as a free sample against shipping cost only, so the verification any bulk RFID tag key fob supplier discussion should start with is a sample on your own reader, at your own MOQ, before tooling money moves.

 

Both routes into the checklist are supported. If you already have a drawing, sample or written specification, the OEM route takes it as given and produces to it. If you know the application but not the chip, the ODM route works backwards from reader model, environment, read distance requirement and order volume. Everything that can be varied is open: frequency, chip, material, mold, printing, encoding and packaging. A supplier willing to put MOQ and lead time in writing at that stage rather than after tooling is telling you something useful about how the rest of the order will go: standard housings with logo and encoding start at 500 pieces on a 7–15 working day production window.

 

Frequently Asked Questions

 

Before you commit tooling money, commit a sample. Send the checklist above with your reader model and a photo of the mounting position, and request a free sample from the China factory that will run the order, tested on your actual doors rather than a desk: start a specification review with our engineering team. A 13.56MHz RFID tag key fob wholesale order with your own custom logo is a two-week decision; a batch that fails on site is a two-month one.