Nuisance Tripping Investigation on a Shared Circuit

Purpose

A breaker that opens under load is usually doing exactly what it was bought to do, and the shop that replaces it without measuring the load has sold a part that will trip again in a week. This procedure puts a number on what the circuit actually carries before anyone decides whether the problem is the load, the breaker, or the connection feeding it.

It also separates the three things a customer calls a nuisance trip. A thermal overload takes minutes and will not reset until the breaker cools. A short circuit or ground fault opens instantly and violently and belongs to the ground-fault procedure. Electronic ground-fault and arc-fault devices trip on something other than current entirely. Treating all three as one complaint is how a circuit gets a new breaker, then a new receptacle, then a new breaker again.

Scope

Covers a standard thermal-magnetic breaker opening repeatedly on a 15 A or 20 A branch circuit serving more than one room or more than one appliance in a dwelling or small commercial occupancy, through to a load finding, a breaker finding or a connection finding.

Does not cover ground-fault or arc-fault device operation, which the GFCI and AFCI Nuisance Trip Service Visit SOP owns, or a hard fault to ground, which the Ground Fault Location SOP owns. Does not cover feeder or main breaker trips, load calculation for a service upgrade, or panel replacement, each of which has its own SOP.

Roles and handoffs

Procedure

1. Characterize the trip from how it resets, before you open anything. Ask whether the breaker resets immediately or only after several minutes, whether the trip is silent or accompanied by a bang, and how long the circuit runs before it goes. Acceptance: a written classification as thermal (minutes of run time, will not reset until cool), magnetic (instant, at the moment something is switched on), or electronic device operation. Wrong looks like recording "keeps tripping." Stop rule: a magnetic-signature trip leaves this procedure for the Ground Fault Location SOP, and a device with a test button leaves for the GFCI and AFCI SOP. Hazard: none at this step, it is an interview.

2. Inventory everything on the circuit, not everything in the room. Establish which outlets, fixtures and hardwired loads the breaker actually feeds. Acceptance: a written list of every point on the circuit, confirmed by turning the breaker off and walking the space rather than by the panel schedule. Wrong looks like trusting a handwritten schedule, which is right about as often as it is wrong. Stop rule: an inventory you cannot complete in the time booked goes to the Circuit Tracing and Mapping SOP as its own visit. Hazard: none beyond normal site work at this step, since the circuit is off and no enclosure is open.

3. Measure the load profile over a representative period, not at one moment. Clamp a true-RMS meter or a logging clamp on the circuit conductor at the panel and record current while the customer runs the combination that trips it. Acceptance: a recorded peak current in amperes with the duration it was held and the loads that produced it. Wrong looks like a single spot reading with the kettle off. Stop rule: no reproducible peak after a full cycle of normal use means this is intermittent, and it moves to the Intermittent Fault Investigation and Monitoring SOP with the logger left in place. Hazard: the deadfront is off and the bus is live while you clamp, so wear the arc-rated clothing and face protection your program assigns per NFPA 70E-2021, 130.5 and 130.7, clamp one conductor at a time, and keep the free hand out of the enclosure.

4. Compare the measured peak against the breaker rating and the continuous-load rule. A load expected to run three hours or more is a continuous load, and NEC 210.20(A) requires the overcurrent device to be rated for the noncontinuous load plus 125 percent of the continuous load, which puts the working ceiling on a 20 A breaker at 16 A continuous and 20 A for anything shorter. Acceptance: the peak, its duration, and which of the two ceilings governs, stated together. Wrong looks like comparing a 40-minute peak against the 16 A continuous figure and condemning a load that is not continuous. Stop rule: a peak at or above the governing ceiling makes this a load problem and step 5 does not run; the fix is redistribution or a new circuit, not a bigger breaker on the same conductor. Hazard: none at this step, it is arithmetic done at the truck.

5. Check the breaker's own connections thermally, under load, before condemning it. With the circuit carrying its measured peak and the deadfront off, compare the temperature of the suspect breaker's line-side connection against an identically loaded breaker in the same panel. Acceptance: the suspect connection within a few degrees of the reference connection at the same current. Wrong looks like tens of degrees above the reference, which is a resistive joint dissipating heat into the breaker's own thermal element and tripping it early on a load it should carry. Stop rule: a rise of that size stops the load conversation entirely, because no amount of redistribution fixes a loose stab. Hazard: an infrared reading with the deadfront off is energized work at full arc-flash exposure, so it is taken from outside the panel with no part of the instrument or the hand entering the enclosure, and the thresholds and reference-component method belong to the Infrared Scan of a Panel SOP.

6. Account for the panel's ambient before blaming the breaker. Record the air temperature at the panel and note whether the adjacent positions are heavily loaded. Acceptance: the ambient written down alongside the load profile. Wrong looks like ignoring that molded-case breakers are calibrated at a standard ambient stated in UL 489, so the same panel in an unconditioned garage in August trips earlier than it does in February at the same current. Stop rule: a circuit that only trips in hot weather at a peak below the governing ceiling is an ambient finding, and the fix is load redistribution or relocation rather than a replacement breaker. Hazard: none at this step, it is a reading taken outside the enclosure.

7. Replace the breaker only after the connection and the bus have been examined dead. Where the finding is the breaker or its connection, open the main, prove dead on a known live source before and after per NFPA 70E-2021, 120.5 with work practices at 29 CFR 1910.333(b)(2), remove the breaker and inspect the bus stab. Acceptance: a stab with light discoloration only, no pitting and no metal loss, cleaned and fitted with a breaker listed for that panel. Wrong looks like snapping a new breaker onto a pitted stab, which repeats the failure with a part on the invoice. Stop rule: pitting, metal loss or a stab that has lost spring tension ends the repair and becomes a panel replacement conversation. Hazard: the main lugs and service conductors stay live with the main open, so nothing above the main gets touched, and the breaker is removed by rocking it off the stab rather than by pulling straight out.

8. Put the circuit back and prove it against the sequence that failed. Restore power and re-run the exact load combination from step 3 for at least the duration that used to trip it. Acceptance: the same peak current within a couple of amperes, held for that duration with no trip, and the step 5 thermal comparison repeated at that load and now within a few degrees of the reference. Wrong looks like closing the breaker, hearing nothing and leaving. Stop rule: a trip during the re-run means the finding was wrong and the procedure restarts at step 3, not at the parts counter. Hazard: this is where the energy goes back with the customer present, so clear the space in front of the panel, stand to the hinge side, close with the flat of the hand, and confirm the deadfront is back on before the load sequence runs.

The record this produces

One investigation record per circuit: the step 1 classification; the step 2 point-by-point inventory; the step 3 peak current with its duration and the loads that made it; the governing ceiling from step 4 and which of the two it was; the step 5 thermal comparison with both temperatures and the current they were taken at; the ambient; the bus stab condition; and the step 8 re-run peak, duration and thermal pair.

The office reads the peak and the ceiling to explain a quote for a new circuit to somebody who thinks they are being upsold. The next tech reads the inventory, because it is the only correct list of that circuit anyone has ever written. The customer reads the before and after, which is what makes a repair believable.

Worked pass: 20 A circuit serving a kitchen counter and a dining alcove, trips two or three times a week

Step 1: it trips about seven minutes into breakfast, will not reset for several minutes, and there is no bang. Thermal signature, so it stays in this procedure.

Step 2: breaker off and the space walked. Eleven points, of which the customer knew about six. The alcove receptacle feeding a chest freezer was not on the schedule at all.

Step 3: logging clamp on the circuit conductor. Running the toaster oven, the microwave and the freezer together gives a peak of 18.4 A held for 40 minutes before the breaker opens.

Step 4: 40 minutes is not three hours, so the load is noncontinuous and the 20 A rating governs rather than the 16 A continuous figure, which is 20 times 0.8. At 18.4 A against a 20 A ceiling the circuit is inside its rating, so this is not a load problem and step 5 runs.

Step 5 fails and takes its stop rule. At 18.4 A the suspect breaker's line-side connection reads 71 C while an identically loaded 20 A breaker three positions away reads 37 C at the same moment. That is 71 minus 37, or a 34 C rise over the reference, which is far past a few degrees. The load conversation stops here.

Step 6: ambient at the panel is 24 C, well inside normal, so ambient is not carrying any of this. Recorded anyway.

Step 7: main open, panel proved dead on a known live source before and after. The breaker comes off and the bus stab shows brown discoloration but no pitting and no metal loss, and the stab still has spring tension. Cleaned, and a breaker listed for that panel is fitted.

Step 8: same three loads, same order. Peak 18.3 A, within a couple of amperes of the 18.4 A recorded before, held 55 minutes with no trip. The thermal comparison at that load reads 41 C on the repaired position against 38 C on the reference, a 3 C difference. Circuit released, with the freezer noted on the schedule and a separate circuit for it quoted on its own merits rather than as the fix.

References

• NEC 210.20(A) and the Article 100 definition of a continuous load, in the edition your authority having jurisdiction has adopted, for the ceilings applied at step 4
• UL 489 for the standard calibration ambient of molded-case circuit breakers referenced at step 6, and the breaker manufacturer's own instructions for panel compatibility at step 7
• 29 CFR 1910.333(b)(2) for work practices, with NFPA 70E-2021, 120.5 for live-dead-live and 130.5 and 130.7 for the arc-flash risk assessment and PPE used with the deadfront off
• See related: GFCI and AFCI Nuisance Trip Service Visit SOP, Infrared Scan of a Panel SOP, Breaker Replacement and Compatibility Verification SOP