Factory firmware that gives every unit an identity
The factory is where every unit gets its bootloader, its first version, its own identity and a locked debug port. We write that flow so a contract manufacturer can run it and you can audit it.
Key characteristics
- Covers
- Programming, factory test, provisioning, traceability
- You provide
- Manufacturer contact, station hardware, PKI or cloud account
- You receive
- Factory image, test firmware, scripts, station procedure
- Every unit
- Its own identity, a locked debug port, a log record
A prototype is programmed by an engineer with a debug probe on a desk. A production unit is programmed by a station on a contract manufacturer’s line, in seconds, by an operator who has never seen your code. Everything the engineer did by hand has to become a procedure: which image, which keys, which tests, what to do with a unit that fails, and what gets recorded. If that procedure is wrong, every unit carries the mistake, and some mistakes, like an open debug port or a key shared across the fleet, cannot be fixed later with an update.
We write the factory firmware and the procedure around it, and we test both before the first production run.
What the factory flow does
- Program the boot chain. The bootloader, with your public verification key built in, and the first application image at the shipping version. This decides what the unit will accept for the rest of its life: with MCUboot, for example, every later image must carry a signature that matches the key in the bootloader.
- Run factory test firmware. A test image, or a test mode, checks every part a customer depends on: sensors against known references, radio transmit and receive, flash, buttons, lights and battery measurement. It is fast, and it gives a clear pass or fail.
- Provision identity. Each unit gets its own serial number, its own device key or certificate, and any network credentials it needs. Never a shared fleet password.
- Lock the unit. The debug port is locked, readout protection is enabled where the chip offers it, and factory commands are disabled.
- Record. The serial number, board revision, firmware version, test results, station, time and the public half of the unit’s identity go into a log your systems can query.
- Handle failures. A unit that fails goes to rework, retest or scrap by a rule written in advance, not by a decision made on the line.
Per-device identity and keys
A device identity lets your cloud know that a message came from one particular unit, and lets you revoke a single stolen or cloned unit without touching the rest. The strongest designs generate the device’s private key on the device itself, or in a secure element, so it never exists anywhere else; the station only collects the public key or a certificate request and has it signed by your certificate authority. Where the hardware cannot do that, keys come from a provisioning service in your own account over a controlled link, and the station does not keep them.
Regulators have started to ask for this. The UK’s product security regime, in effect since April 29, 2024, requires passwords on consumer connectable products to be unique per product or defined by the user. For products in its scope, the EU Cyber Resilience Act asks for protection from unauthorized access through mechanisms such as authentication and identity management, and for products designed to limit their attack surfaces, including external interfaces; its main obligations apply from December 11, 2027. This is not legal advice; firmware security covers the wider picture.
Debug lock without losing diagnosis
Locking the debug port is usually the last step at the station, because on some chips the lock is permanent. That ordering, and how a returned unit will be diagnosed once it is locked (through stored logs and a version and health readout, not a debug probe), are written down and tested before the first production run. The secure boot checklist includes the lockdown checks: debug access attempted on a locked unit and refused, readout protection on, factory commands gone.
Traceability
When a field problem appears, the first question is which units are affected. Traceability answers it. Every unit’s serial number links to its board revision, the firmware version it left the factory with, its test results, and the date and station it was built on. Combined with the version reports devices send after each update, you can tell which units run what, today. We design the record format with you and make sure your manufacturer can produce it in a form you can import.
Where it fits in production gates
Our free production readiness checklist splits the run-up to production into EVT, DVT and PVT gates. Factory firmware belongs mostly to PVT, the gate that asks whether the factory can build the product: the programming procedure written and timed, factory test checking every part a customer depends on, unique identity provisioned and recorded for each unit, the debug port locked, and a defined path for failed units. The last gate also asks that support can read version and health from a returned unit, which the factory flow makes possible.
Example (fictional). Loamwick’s requirement FW-006 asks that factory test finish in under 60 seconds per LW-1 probe, verified by timing 50 boards on the test fixture. The station programs MCUboot with Loamwick’s public key, the signed application image and the probe’s own identity, then locks the debug port. Because every probe’s record links its serial number to its board revision, Loamwick can list the affected units at once when a known issue touches only one revision, such as Rev C probes reading temperature 0.5 °C high above 40 °C. See the example project.
What you receive
- The factory image and factory test firmware.
- Provisioning scripts that work with your certificate authority or cloud account.
- A station procedure written for the manufacturer’s operators, with timings.
- The log format, with a sample, and the documented failure path.
- Test records from dry runs of the whole flow on pre-production units.
Where to start
If your first production run is a few months away, start with a firmware review: we look at your boot chain, keys, debug access and test plan, and tell you in writing what the factory flow needs and what to settle with your manufacturer now. Request a firmware review, or read more in the prototype to production section of Field Update.
Send us your board and your plan.
A firmware review reads your schematics, code, flash layout and update plan, and ends with a written plan: what to build first, what could go wrong in the field and how updates will reach your devices.
Or email sales@fumbar.com, or call (720) 675-4202.