How to Read a Trip as Evidence

Why this matters

A trip is a dated, timed measurement taken by a device that was there when you were not. Almost all of that measurement is perishable, and the most perishable parts are destroyed by the two things that happen first: the customer resets it, and everything cools off. A tech who arrives and immediately starts testing has usually already lost the half of the evidence that could not be recovered, and kept the half that would still be there next week.

These steps are ordered by how fast the evidence disappears once the trip has happened, fastest first. That is a different order from the one that feels natural, which is to start with the meter. The meter measures things that will still be measurable tomorrow.

Before any of it: the gate on working in the panel

Every step below that happens with a cover off is live work. 29 CFR 1910.333(a)(1) requires conductors and parts to be de-energized before employees work on or near them unless the employer can demonstrate that de-energizing introduces additional or increased hazards or is infeasible due to equipment design or operational limitations, and that demonstration belongs to the employer's program, not to a judgment made at the panel. Where live work is justified, the approach boundary and arc-rated PPE come from the risk assessment at NFPA 70E-2021 130.5 and the PPE tables at 130.7, in the edition your employer's electrical safety program has adopted, since NFPA 70E binds through that program or a contract and not on its own. For the isolation half, the fork matters: 29 CFR 1910.147 excludes exposure to electrical hazards from work on conductors and equipment in electric utilization installations at (a)(1)(ii)(C), so panel work runs under 29 CFR 1910.333(b)(2) in general industry and 29 CFR 1926.417 on construction work, with live-dead-live proving at NFPA 70E-2021 120.5.

Step 1: Capture heat, because it is gone in minutes

Once the device opens, current stops, and every thermal clue in the enclosure begins decaying immediately. A lug that was running warm under load is at ambient well before most techs arrive.

Where you get to a fresh trip - you are on site, or the customer calls while it is still open - a thermal scan of the device body, both terminations and the adjacent poles, taken with the cover off for only as long as the scan takes and under the gate above, is the one reading you can never get later.

What it costs when it is gone: the difference between a device operating on circuit current and a device operating on heat arriving from a bad joint next to it. Those two produce identical trip records and different repairs.

Step 2: Record the as-found handle and indicator, because one reset erases it

Look before you touch. A tripped molded-case breaker usually sits in a middle position rather than fully off, which distinguishes a trip from somebody switching the circuit off. Devices with added functions often carry a separate indicator for what operated - a ground-fault or arc-fault mechanism will frequently latch differently or show a flag - and the reset behaviour itself is data.

Photograph the panel with the directory visible before restoring anything.

What it costs when it is gone: the ability to distinguish a trip from a switch-off, and on a multifunction device, the ability to tell an overcurrent event from a ground-fault or arc-fault event without recreating the whole thing.

Step 3: Reconstruct what was running, before the memory decays

Ask within a day if you can. What was on, what had just started, what else was running at the same time, what the weather was doing, whether anybody had been working in the building. Ask for the times, not the story.

Note that this is partly recoverable later - a customer can look at their own schedule, a facility can pull a log - which is why it sits below the two steps that are not recoverable at all.

What it costs when it is gone: the correlation. Without it, a trip is an event with no cause attached, and you are reduced to guessing which of the loads on the circuit was involved.

Step 4: Establish time from energization to trip, or plan to buy it back

If nobody knows, you can only recover this by re-running the event under observation, which costs a return visit and sometimes a full day of load time. Everything below this step survives until somebody works on the circuit; this one does not survive at all without either a record or a rerun.

Ask for it in bands, because that is all you need: immediately, within a minute, within an hour, or after several hours of running.

What it costs when it is gone: an entire visit, and often the customer's patience, since the rerun is the trip they called you to prevent.

Step 5: Name which element operated, from the timing

This is derived from step 4, so it lives exactly as long as step 4's record does. Three bands and what each one implicates:

• A fraction of a cycle, every time, at energization. The magnetic element, which needs a multiple of rating - commonly in the region of ten times rating for a standard molded-case breaker, though the multiple is a property of that product's published curve rather than a constant. That is fault territory.
• Minutes to about an hour under a steady load. The thermal element, which integrates heat. That is load or ambient territory. Note the continuous-load limit here: a standard branch device is not listed to carry more than 80 percent of its rating for three hours or more unless the assembly is listed for 100 percent operation.
• At any time, at a current far below rating. Not an overcurrent element at all, but a ground-fault or arc-fault function. A Class A ground-fault circuit interrupter operates in a nominal 4 to 6 mA band under UL 943, the listing standard the device was tested to in the edition it was listed under, which binds through the listing and through the adopted NEC's requirement that the device be listed.

What it costs when it is gone: you test the wrong thing. Insulation resistance testing a circuit that tripped thermally at 40 minutes wastes an hour and finds nothing, because nothing is wrong with the insulation.

Step 6: Take the de-energized electrical evidence

This survives until somebody works on the circuit, which is why it goes here rather than first. De-energized, locked out and proved dead under the standards named above: continuity, terminations retorqued to the manufacturer's value, conductor and insulation condition at every point where the run is disturbed or flexed, and where the timing implicated a fault, an insulation resistance test.

An insulation resistance test is an instruction that creates its own hazard: the instrument applies hundreds of volts DC and charges the cable's capacitance, so disconnect electronic loads and any device that cannot take the test voltage before you start, keep everyone off the far end, and use the instrument's discharge function and confirm the conductors are at zero before touching them.

What it costs when it is gone: nothing, if the circuit is untouched. It is the one step you can reliably come back for, which is exactly why it should not be the one you do first.

Step 7: Write the record that makes the next trip readable

The record does not decay at all. It also never gets created retroactively, which is the whole problem. Date, time, device and circuit, as-found handle position, time from energization to trip, measured current and ambient, what was running, and what you changed.

What it costs when it is gone: the third trip is diagnosed from scratch by the third tech, which is how a two-hour problem becomes three visits.

Worked example: three trips in ten days, all reset before anyone looked

A 20 A branch circuit at a small commercial space. Three trips in ten days, each reset by staff within minutes. Nothing found on two previous visits.

Steps 1 and 2 are already gone - the panel has been closed and reset for three hours, and nobody photographed anything. That loss is the finding for later; it is not recoverable now.

Step 3, taken from the customer's own daily log rather than memory: all three trips landed in the afternoon, and all three were within about 12 to 20 minutes of a second piece of equipment being switched on alongside the one that runs all day.

Step 4 has to be bought back, so it is scheduled: energize normally, clamp on the circuit conductor, stopwatch running, under the live-work gate above.

The rerun, with the numbers carried through:

• Baseline with the all-day equipment only: 11.5 A. Against the 20 A device that is 58 percent of rating, and it is under the 16.0 A continuous limit that 80 percent of 20 A gives.
• Second machine started: current steps to 17.8 A. That is 89 percent of rating and 1.8 A above the 16.0 A continuous limit, or about 11 percent over it.
• Trip at 24 minutes.

Step 5 reads that straight off the bands above: 24 minutes under a steady load is the thermal element. Not a fault, not a ground-fault function, not inrush. The device integrated heat from a load that sat 11 percent over what it is listed to carry continuously, and it opened.

Step 6 is still worth doing and finds nothing, which is the correct result and confirms the reading rather than wasting the visit. Insulation is sound, terminations are tight, no damage anywhere.

The answer is a load answer, not a device answer. Two machines on one 20 A circuit put it above its continuous rating whenever both run. Moving the second machine to its own circuit is a design change governed by conductor ampacity under NEC Article 310 and the branch-circuit rules of Article 210, in the edition your authority having jurisdiction has adopted, and the sibling article on what an overcurrent device is protecting explains why the answer is never a larger breaker on the same conductor.

The failure mode if step 5 is skipped: the previous two visits both went straight to step 6, found sound insulation, and concluded the breaker was weak. A replacement device would have behaved identically, because 17.8 A on a 20 A breaker for 24 minutes is inside what any correctly built device does.

What would have changed the conclusion: a trip at energization rather than at 24 minutes. Same load story, same customer report, and the correlation with the second machine would then be a red herring pointing at a fault in that machine's supply rather than at circuit loading.

How to verify you read it correctly

• Predict the next trip before it happens. If your reading is right, you can state the condition and the rough time-to-trip in advance. Here that is: both machines running, roughly 20 to 30 minutes. A reading that cannot make a prediction has not identified a mechanism.
• Check that your named element matches your named cause. A thermal cause with an instantaneous trip, or a fault cause with a 40-minute trip, is a contradiction and one of the two is wrong.
• Confirm the fix removes the condition, not the report. After the second machine moves, the original circuit should sit back near 58 percent of rating and hold indefinitely. If it still trips, the correlation was coincidental and steps 3 through 5 need re-running.

References

• 29 CFR 1910.333(a)(1) and (b)(2), safety-related work practices for electrical work; 29 CFR 1926.417 for the construction counterpart
• 29 CFR 1910.147(a)(1)(ii)(C), the carve-out for electric utilization equipment
• NFPA 70E-2021, 120.5, 130.5 and 130.7, as adopted by your employer's electrical safety program
• UL 943 and UL 489, the listing standards for ground-fault circuit interrupters and molded-case circuit breakers, in the editions the installed devices were listed under
• See related: Why a Breaker That Trips Is Usually Not the Fault; What an Overcurrent Device Is Actually Protecting