What is Selective Coordination?

A fault happens on one branch circuit in a 30-story building. Which breaker opens?

If you’ve designed the system well, only the branch breaker for that specific circuit opens. Everyone in the rest of the building keeps their lights, their HVAC, their elevators. The maintenance team investigates one outlet’s worth of failure, not an entire floor’s.

If you haven’t designed the system well, the main breaker at the service entrance opens — and 30 floors go dark.

Selective coordination is the discipline of making sure the first answer is what happens, not the second.

The definition

A system of overcurrent protective devices is selectively coordinated if, for any overcurrent condition, only the device immediately upstream of the fault opens — leaving every other device in the system intact.

That’s it. One sentence. Most of the complexity of the discipline comes from proving that the property holds across every possible fault current, not just the worst-case symmetrical short circuit.

Why the code cares

The National Electrical Code makes selective coordination mandatory for three specific kinds of power systems:

• Article 700 — Emergency Systems. Egress lighting, fire alarms, smoke control, life-safety equipment. A breaker tripping the wrong level here can kill people in a real emergency.
• Article 701 — Legally Required Standby. Systems whose failure would create a hazard or hamper rescue operations — sewage pumps in flood districts, ventilation in industrial process buildings, communications for first responders.
• Article 708 — Critical Operations Power Systems (COPS). A more recent classification covering facilities the government deems critical to public safety: certain hospitals, data centers, telecom hubs, military installations.

For systems in those categories, the engineer of record must produce a coordination study demonstrating that every series pair stays selective across the full range of available overcurrents — from overload all the way up to the available fault current (and the full range of device opening times). That is how the NEC defines selective coordination. Some authorities having jurisdiction accept a practical simplification — checking coordination only down to 0.1 second rather than all the way into the instantaneous region — but that 0.1 s cutoff is a softening of the requirement, not the baseline.

Why everyone else cares

Outside the code-required systems, selective coordination isn’t legally mandatory. It is, in practice, the difference between a tenant complaint and a six-figure outage.

A grocery store with poor coordination loses every freezer in the building because a compressor faulted. A data center with poor coordination loses an entire production rack because a single power supply burned up. A manufacturing line with poor coordination loses the whole shift because one motor stalled.

Whenever continuity-of-service has any economic value, coordination pays for the engineering it took to design it.

What it looks like on a TCC

Put two devices in series — a main upstream, a branch downstream — and plot both curves on the same TCC. Coordinated means: for every current on the x-axis, the downstream curve is below the upstream curve. The downstream device clears faster at every fault level.

Visually, the two curves never touch. There’s daylight between them everywhere.

Notice three properties of the coordinated plot above:

1. The branch (orange) curve is always below the main (blue) curve.
2. The main breaker’s instantaneous is OFF — its short-time delay carries the curve straight to the right edge of the plot. This buys the downstream breaker time to clear faults at every current, including the maximum bolted fault.
3. There’s visible separation between the two curves at every point. A coordination study will demand a specific minimum separation (often 0.1 s) — but the principle is “downstream below upstream, everywhere.”

Where it breaks down

Selective coordination breaks down in three places, and you’ll spend most of Module 2 learning to spot each one on a plot:

• Inverse-time overlap in the long-time region. Two breakers with similar thermal trip behavior can swap order at low overload currents.
• Instantaneous-on-instantaneous. If both devices have instantaneous trips, and the available fault current exceeds both instantaneous pickup levels, both devices race — and neither one guarantees it’ll be first.
• Short-time on short-time without an intentional delay step. Two LVPCBs with S elements need their S delays staggered, or the upstream LVPCB might pick up its S element before the downstream one finishes opening.

Lesson 5 puts you in front of all three failure modes with a live two-device plot.

What’s next

In Lesson 5 you’ll get a draggable two-device plot and a deliberately miscoordinated starting point — you’ll fix it by shifting the upstream device’s settings until the two curves no longer cross.

In Lesson 6 you’ll do the same exercise across a three-device cascade — utility → main → feeder → branch — which is the real working pattern for industrial and commercial coordination studies.

Then in Module 3 we’ll add cable and transformer damage curves to the plot so coordination becomes not just which device opens first but does the right device open before something burns up.