The Hidden Cost of a Component Change
Why hardware companies discover certification failures too late.
A certification is a statement about one specific physical configuration. It says that a particular device, built a particular way, was measured against a particular requirement and found to be within limits. Change the configuration and the statement may no longer hold — even when the firmware, the user-facing behaviour, and the bill of materials description all look unchanged.
Component substitutions are the clearest example. A capacitor goes end-of-life. A second-source RF module is approved by procurement to protect the line. A connector is swapped for a pin-compatible equivalent. On paper these are like-for-like. Physically, they can move the very properties a certification measured.
Why the cost stays hidden
The difficulty is rarely knowledge. Most engineers know that physical changes can affect compliance. The difficulty is timing and visibility. The change happens quietly — in a CAD revision, a procurement decision, or a single commit — and the consequence only becomes visible weeks later, in a test report, after downstream work has already been built on the assumption that the design was certifiable.
By then the change is no longer one decision. It is embedded in tooling, in firmware that targets the new part, in marketing dates, and in commitments to customers. The cost of unwinding it is far larger than the cost of the original substitution.
The cheapest re-test is the one you never have to book.
The changes that most often invalidate a certification
Regulatory certifications for connected hardware are sensitive to the parts of a design most likely to change late in a programme. The recurring offenders are physical rather than logical:
- Second-sourcing or substituting an RF module or front-end
- Moving, re-orienting, or re-tuning an antenna
- Changing trace geometry, ground plane, or shielding near the radio
- Raising conducted output power or antenna gain
- Swapping power components that change conducted or radiated emissions
Each of these can move radiated power, spurious emissions, or occupied bandwidth relative to a regulatory limit. None of them necessarily touch the firmware, so none of them reliably trip a standard code review or a functional test.
Making the impact visible at design time
The remedy is not more testing at the end. It is making the link between a change and its certification impact explicit at the moment the change is made — while the engineer still holds the context needed to act on it.
Concretely: when a component is substituted, the system that holds the certification record should already be able to surface which requirement governs the affected behaviour, what the certified margin was, and whether the new part plausibly threatens it. That turns a future lab surprise into a present-day engineering decision — one that can be reviewed, accepted, or escalated on its own terms.
A practical discipline
Teams that avoid late failures tend to share one habit: every change that touches a certified property carries a short, traceable note before it lands. The note answers three questions — what changed, which requirement it touches, and what the available evidence says. Nothing more. It is a small discipline, and it is also the difference between discovering a problem on a screen and discovering it in a chamber.
Crado provides engineering decision support and traceability tooling. Final certification and regulatory decisions remain the responsibility of qualified engineers, certification bodies, and relevant authorities.
- FCC Rules — Title 47 CFR Part 15 — U.S. Electronic Code of Federal Regulations
- Radio Equipment Directive 2014/53/EU — EUR-Lex