The 125% continuous-load rule

What makes a load continuous, why the factor is applied twice, and why a 40 A charger needs a 50 A breaker but a 40 A cooktop does not.

Updated

Two circuits, both drawing 40 amperes. One needs a 50 A breaker and 8 AWG conductors; the other is fine on a 40 A breaker and 8 AWG. Nothing about the current is different. The difference is how long it stays there.

The definition

The NEC defines a continuous load as one where the maximum current is expected to continue for three hours or more. That is the whole test — a duration, not a load type.

Three hours is a thermal number rather than an arbitrary one. A conductor carrying current heats up and reaches a steady temperature; the time constant for a typical circuit puts that equilibrium roughly in the three-hour range. A load that stops before then never reaches its final temperature. A load that runs past it does, and everything downstream — conductor insulation, breaker calibration, terminal connections — is living at that temperature indefinitely.

So the question is never “is this a big load” or “is this an important load”. It is: will it sit at its maximum for three hours?

Working through the everyday cases:

  • Continuous. Commercial and retail lighting on all day. EV charging, which the code treats as continuous by rule rather than by judgement. Sign and parking-lot lighting. Server and network equipment. Process equipment on a shift.
  • Not continuous. Domestic cooking appliances. Receptacle circuits in a dwelling. Laundry equipment. A domestic dishwasher. Anything cycling on a thermostat with a duty cycle well under three hours.
  • Judgement required. A motor on a production line, which depends entirely on the process. Domestic lighting, which is usually not continuous but can be in a room that is lit all evening. Where you genuinely cannot tell, treating it as continuous is the conservative answer and costs one size.

Why 125%, and why twice

The factor is 125%, which is the reciprocal of 80%. The code is saying the same thing in two directions: a circuit carrying a continuous load must not be loaded beyond 80% of its rating, which is identical to sizing it at 125% of the load.

The part people miss is that it applies in two separate places, under two separate rules.

The conductor — 210.19(A)(1) for branch circuits, 215.2(A)(1) for feeders — must have an allowable ampacity of at least 125% of the continuous load.

The overcurrent device — 210.20(A) and 215.3 — must have a rating of at least 125% of the continuous load.

Both, not either. They tend to land together because conductors and breakers are sized from the same design current, which is why treating it as one rule usually produces the right answer. It stops being reliable the moment a conductor is upsized for voltage drop, or a termination temperature limit binds, or a derating factor applies — and then a circuit can end up with a conductor sized properly and a device that is not, or the reverse.

A third rule, 110.14(C), then caps everything at the termination temperature rating, and it is applied to the result of the 125% calculation, not before it. There is a separate article on termination temperature ratings.

The two circuits

Return to the 40 A pair.

The EV charger. 625.41 requires EVSE loads to be treated as continuous, no judgement needed. So 40 × 1.25 = 50 A of required capacity. 50 A is a standard rating in the 240.6(A) list, so the answer is a 50 A breaker, and the conductor must be good for 50 A at its termination rating — 8 AWG copper at 75 °C, exactly. No rounding, no headroom, and if the run is long enough that voltage drop pushes the conductor to 6 AWG, the breaker stays at 50 A.

The cooktop. A domestic cooking appliance runs in bursts through a meal. Not continuous, so no factor applies. The required capacity is the load itself, 40 A, which is also a standard rating, so a 40 A breaker and a conductor good for 40 A.

Same current, one device size apart, decided entirely by the duration question. Run either through the breaker size calculator and the load type is the only input that changes.

Where it lands exactly, and why that is uncomfortable

Notice what happened in the EV case: 40 × 1.25 = 50, and 50 is a standard breaker size. The required capacity sits precisely on the device rating, with nothing spare.

That is compliant and it is also worth pausing on. Add anything to that circuit later and it no longer complies. This is not a defect in the rule — 125% is the margin, and it is already in the number — but it does mean the tidy-looking answers are the ones with the least room. The messy ones, where the calculation lands at 47.5 A and rounds up to 50, are the ones with slack.

The 100% rated exception

There is a way out, and it is rarely available. 210.20(A) and 215.3 both permit the overcurrent device to be sized at 100% of a continuous load where the assembly — the breaker and the enclosure it sits in, as a listed combination — is rated for continuous operation at 100% of its rating.

Three things to know about it. These assemblies exist mostly at larger frame sizes and in commercial switchgear, so it is not an option on a residential panel. The listing is for the combination, not the breaker alone, so a 100% rated breaker in an ordinary panelboard does not qualify. And the exception applies to the device only — the conductor rule is unchanged, and conductors still need 125%.

The mistakes worth naming

Applying the factor to the wrong current. For a motor, the 125% in 430.22 runs on the table full-load current from Article 430, not on the nameplate and not on a measured figure. Different rule, different starting number, same-looking multiplier.

Applying it twice. Where a rule already includes the continuous allowance — the MCA printed on HVAC equipment, for instance, which the manufacturer has already computed — multiplying again oversizes the circuit. The nameplate figure is the answer, not the input.

Applying it to the conductor but not the device, or the reverse. Two rules. Check both.

Rounding a ceiling upward. 125% produces a minimum, so it rounds up. Elsewhere in the code the same-shaped arithmetic produces a maximum that must round down — motor branch-circuit protection, and the MOCP on air-conditioning equipment. Getting these confused is the single most common error in this territory, and there is a separate article on rounding up and rounding down.

The short version

Three hours at maximum current makes a load continuous. Multiply by 1.25 and apply it to both the conductor and the overcurrent device. Cap the result at the termination temperature rating. And when the arithmetic lands exactly on a standard size, notice that you have no headroom at all.