Overcurrent device coordination is frequently assumed rather than demonstrated. Design documentation shows one protection philosophy, the as-installed devices reflect procurement substitutions and late-stage changes, and the coordination study — if it exists — often references neither accurately.

The protective intent is straightforward enough: the device closest to the fault clears it selectively, leaving upstream circuits unaffected. In a critical power distribution environment, failure of that selectivity means a fault deep in the secondary distribution takes out a section that should have stayed live. The engineering principle is not the problem. The gap is in verification — whether anyone has actually confirmed that the installed devices, at the available fault level, achieve selective operation across the trip current range.

What coordination evidence should look like

A coordination study plots the time-current characteristics of the overcurrent devices in the protection hierarchy, demonstrating that the downstream device clears before the upstream device operates within the fault current range present at that point. That study is only meaningful to the extent it references the devices actually installed. Not the devices on the original specification, not the devices on the purchase order three months before practical completion — the devices in the installed condition at the point of review.

Relay settings — the actual programmed pickup and time-delay values, not factory defaults — need to be documented and cross-referenced against the study. Manufacturer characteristic data for the installed devices should be present in the package. When those three elements align, the coordination claim can be evaluated. When any one is missing or inconsistent, the study is not evidence of the installed condition.

Common gaps in submitted packages

Procurement substitution is the most frequent source of misalignment. A device is substituted mid-build — different manufacturer, different frame size, different trip unit characteristic — and no re-coordination check is triggered. The original study remains in the documentation package, looks complete, and doesn't reflect the installed condition. This pattern is routine on medium to large projects, not exceptional.

Substitution without re-study

On an anonymised data centre pre-energisation review, a coordination study was submitted covering the primary protection hierarchy. Cross-referencing against the as-installed schedule identified three device substitutions made during procurement — different manufacturer, different trip characteristic — none of which had triggered a re-study. The submitted study remained in the package unrevised.

The substitutions may or may not have affected selectivity in the relevant fault current range. That question couldn't be resolved from the documentation as submitted — it required re-analysis against the installed devices' actual characteristics.

Final distribution boards

Panels serving UPS outputs, precision cooling units, and PDU feeds frequently have no coordination data submitted against them at all. The implied assumption — that devices selected by rating alone coordinate across the available fault level — may hold for straightforward configurations, but doesn't warrant acceptance without verification. The available fault level at LV distribution boards in large data centre installations can be high enough that the characteristics of RCBOs, MCBs, and moulded-case devices diverge in ways that matter.

At minimum, the available fault level at the board should be documented, the devices confirmed as rated for it, and the time-current characteristic comparison should cover the fault current range actually present. Where that analysis isn't in the package, the coordination at those boards hasn't been demonstrated — it's been assumed.

What the review is looking for

Coordination failures don't present until a fault occurs. A distribution system can operate without incident for an extended period with inadequate selectivity and produce no observable indication of the gap. The problem surfaces when a fault event exposes the protection coordination as insufficient, at which point the scope of impact is determined by what the protection scheme actually does rather than what the design intended.

Verification sequence for review: Is a coordination study present? Does it reference the devices currently installed, or a prior procurement schedule? Were substitutions during build reflected in a re-study? Do documented relay settings match the study? For final distribution boards, are available fault levels documented and devices rated accordingly? Where any of these questions can't be answered from the submitted package, the coordination status is unverified.

The status is binary. Either the study covers the installed condition and demonstrates selective operation across the relevant fault current range, or it doesn't. Plausible alignment between the study and the current schedule isn't the same as demonstrated coordination. At energisation, the distinction matters.