The Recurring Call That Was Really a Tenant Behaviour Problem
Why this matters
The fifth trip to the same unit is where shops lose money and managers lose patience, and it is also where the diagnosis quietly stops being technical. Somebody says "the tenant is doing something," everyone relaxes, the ticket closes, and nine weeks later you are back. On a managed property that shortcut is expensive twice: you have absorbed five visits, and you have handed the manager a finding they cannot act on without accusing a resident, which most managers will simply decline to do. This is one such case, followed from the first repeat call to the confirmed cause, including the two hypotheses that looked right and were not.
The safety call that comes before any of this
A protective device that keeps operating is doing its job, and the wrong first move is to reset it again and leave.
Before diagnosis, at the panel: if the breaker or its surrounding enclosure is warm to the back of the hand, discolored, or smells of burning insulation, do not reset it. Shut off the main disconnect if that can be reached from a dry position, keep people out of the space, and call it in as an immediate finding. If a receptacle on the affected circuit is scorched, warm or loose in its box, that receptacle is out of service now: unplug what is in it, tell the occupant in specific words which outlet not to use and for how long, and tag it.
Where you are going to work inside the panel, de-energize, lock and tag the energy-isolating device, and verify dead with a meter proven on a known live source immediately before and immediately after the reading. For panel work the governing general-industry rule is 29 CFR 1910.333(b)(2), not 1910.147, which excludes electric utilization installations and sends them to Subpart S; the live-dead-live proving step itself is NFPA 70E 120.5. A construction-classified job falls under 29 CFR 1926.417 instead. De-energized and verified is not the same as switched off.
And on this specific case: a portable heater found on a power strip or extension cord gets unplugged before anything else, because that is a fire in progress waiting on a duty cycle. Portable heaters belong plugged directly into a wall receptacle with about three feet of clearance from anything combustible, per long-standing NFPA public education guidance, and no reset of any breaker happens while one is on a strip.
The signal: five calls in fourteen weeks
Unit 3C in a stacked, multi-unit building. Same complaint every time: a breaker trips, power is lost to part of the unit, the tenant calls the manager, the manager dispatches, the tech resets, everything works.
Pulling the call history rather than the last ticket is what made this diagnosable. Five calls in fourteen weeks, roughly one every three weeks. Four of the five reported trip times fell between 5pm and 9pm. One was at 6am. Every call was in the second half of the heating season.
That distribution is the first real clue and it is worth reading carefully. A breaker with an internal defect trips on a schedule that has nothing to do with the clock. A load-driven trip clusters where the load clusters. Four of five in the same four-hour evening band is a pattern; it is not proof, because five events is a small sample and one of them fell outside the band entirely. Note the outlier rather than smoothing it away - the 6am call is the one that eventually stopped a wrong conclusion.
Hypothesis one: the breaker itself
A weak or fatigued breaker will nuisance-trip below its rating, and it is the cheapest thing to eliminate, so it went first. On call two the breaker was replaced with an identical rating.
Nine days later the tenant called again. Same circuit, same evening band. That eliminates the device cleanly: a new breaker exhibiting the identical behavior on the identical circuit is telling you the problem is on the load side or in the wiring, not in the protection.
The cost of learning this the slow way is worth naming. Replacing the breaker felt like a fix, so the ticket closed as resolved, which meant call three arrived at the manager as a new incident rather than the third in a series. Any repeat that gets closed as resolved resets the counter in everybody's head, and the counter is the diagnostic.
Hypothesis two: a fault on a fixed load
The next theory: a deteriorating fixed appliance or a compromised conductor drawing fault current. This is the one most techs would carry longest, because it is the answer that keeps the problem inside the trade.
The test, with the circuit de-energized, locked, tagged and verified: inspect terminations at the panel and at every device on the circuit for heat marking and loose connections, then re-energize and clamp the circuit with only the fixed loads running. Say the circuit is rated 15 A and the clamp reads about 4 A with the fixed loads on and nothing plugged in. Insulation and continuity checks on the branch showed nothing abnormal, no evidence of current on the equipment grounding conductor, no heat marking at any termination.
That is a genuine elimination, not a shrug. A ground fault or a deteriorating termination shows up as heat, as damaged insulation, or as current where there should be none, and none of those were present. But notice the limit of what it proved: it proved the circuit is healthy when only the fixed loads run. It said nothing at all about what happens at 7pm.
Hypothesis three: something building-wide
The building is stacked, units share risers and share a service, and there is a known failure mode where one cause produces symptoms in several units at once. Worth ten minutes to check before spending a day on 3C.
Call history for the two units directly above and below 3C over the same fourteen weeks: zero trips in either. Zero in the rest of the stack. The complaint is confined to one unit, which rules out a shared-cause pattern and rules the utility supply out with it. See related: The Shared System: Multiple Tenants, One Cause.
Two eliminations later the position was uncomfortable and honest: the circuit is sound, the protection is sound, the building is sound, and the unit still trips in the evening. That is where the temptation to write "tenant overloading the circuit" and close the file is strongest.
What was actually true
Instead of writing it, they measured it. A recording clamp meter went on the circuit for 72 hours: about 0.5 hours to place, about 0.5 hours to collect.
The record showed a daytime baseline near 4 A, consistent with the earlier spot reading, and evening peaks reaching just under the 15 A rating and holding there for stretches over 40 minutes, every evening, starting between 5pm and 6pm. On the third evening it crossed and the breaker opened, unattended, with the reading captured.
In the unit: two portable heaters, both on one power strip, on the affected circuit. Both were unplugged on the spot, and the strip came out with them.
So the shortcut answer was right about the mechanism. It was wrong about the cause, and stopping there would have left the actual hazard in place.
Why "tenant behaviour" was the wrong place to stop
The next question is the one that turns this from a blame finding into a repair: why are there two heaters running every evening in a unit with heat.
Room temperature in the living space on the coldest evenings was running several degrees below the thermostat setpoint, and the unit's heat was running nearly continuously without closing the gap. Behind that: a return path badly restricted, and a heating stage that was not coming on. The tenant was not being careless. They were cold, and they had solved it the way people solve it.
That reframes everything about the report. "Tenant is overloading the circuit" is a finding a manager cannot act on and would not want to. "The unit is not holding setpoint on cold evenings, and the occupant is making up the difference with portable heat on a circuit that cannot carry it" is a finding with a work order attached, and it is the same set of facts.
It also explains the outlier. The 6am call sat outside the evening band and did not fit a load-driven theory cleanly, and it was the one event that kept getting waved off. It fits fine once the cause is heat that cannot hold overnight: the heaters were running before dawn on the coldest morning of the series, for the same reason they ran at 7pm.
How this got reported so it produced a fix
Three items, in this order, in writing to the manager the same day:
- The immediate hazard and what was done about it. Two portable heaters on a power strip, removed, with the reason stated in physical terms and the direct-to-wall-receptacle and three-foot-clearance requirement stated for whatever the occupant uses next.
- The measurement. A 72-hour current record showing evening load at the circuit's rating, with the trip captured. Not an opinion about usage, a recording.
- The underlying deficiency. Setpoint not held on cold evenings, restricted return, a heating stage not proving, and the specific work needed.
No price in the hazard report, and no characterization of the tenant. The manager forwarded it to the owner intact, which is the point of writing it that way: a report that requires the manager to rewrite it before an owner can read it usually gets summarized into nothing.
Confirmation
The heating deficiency was corrected. The circuit was left alone, deliberately - nothing was wrong with it.
Eleven weeks of the same heating season followed with zero trips in 3C, against five trips in the fourteen weeks before. The recording meter went back on for 72 hours at the end of that period as a check: evening peaks near 5 A, no sustained approach to the rating.
The economics of the diagnosis are worth keeping. Five reset-and-leave visits at roughly 1.2 hours each consumed about 6 hours across the season and produced no information. The single measurement that ended it cost about 1 hour of tech time plus the meter sitting in place. The lesson is not that instrumentation pays for itself in some ratio - it is that the fourth identical visit is the signal to stop repeating the visit and start recording, and most shops take until the fifth or sixth.
What would have changed the conclusion
If the neighboring units had shown trips too, the confined-to-3C conclusion collapses and the investigation moves to the shared supply and the service, not to anything inside the unit.
If the recording had shown the peak in the middle of the night with the unit unoccupied, the load is fixed equipment on a duty cycle, not occupant behavior, and hypothesis two comes back in a different form.
If the unit had been holding setpoint, the finding really would have stopped at occupant load, and the correct report is then a capacity question - the circuit as built does not support how the space is used - which is an owner decision about the property, not a complaint about a resident.
If this were a condominium rather than a rental, the party who authorizes work inside the unit is the unit owner and the party who authorizes anything upstream of it may be the association, so the same finding splits across two approval paths before anything gets fixed.
References
- 29 CFR 1910.333(b)(2), general-industry rule for electrical work; NFPA 70E 120.5 for the live-dead-live proving sequence
- 29 CFR 1926.417, lockout and tagging of circuits, the construction counterpart
- NFPA public education guidance on portable space heaters: plug directly into a wall receptacle, maintain about three feet of clearance from combustibles
- See related: The Shared System: Multiple Tenants, One Cause
- See related: When the Tenant Is Not Your Customer but Is Your Problem