The number on the table is a ceiling, not a promise

An apprentice pulls a run of #6 THHN, runs a finger down the column, and lands on 75 amps. The math feels finished. The wire is rated, the breaker is picked, the job moves on. Months later a connection runs hot, the insulation smells faintly sweet, and nobody can explain why a conductor that was "good for 75" couldn't hold a 60-amp load on a summer rooftop.

The answer is that the ampacity table never promised 75 amps. It promised 75 amps under one specific set of conditions — a 30°C (86°F) room, a small number of conductors, heat free to escape. Change those conditions and the number changes with them. Learning to adjust it is the difference between a conductor that runs cool for thirty years and one that slowly cooks itself behind a wall.

What ampacity is actually measuring

Ampacity is not a measure of how much current a wire can carry in some absolute sense. Copper will pass enormous current right up until it melts. Ampacity is a measure of heat — specifically, the current at which a conductor's insulation reaches the maximum temperature it was designed to survive.

Every current-carrying conductor has resistance, and resistance turns current into heat. A wire sitting in open air sheds that heat easily. The insulation around it has a thermal limit: 60°C, 75°C, or 90°C, depending on the type. THHN is a 90°C insulation. Table 310.16 simply asks: in a 30°C ambient, inside a raceway, how much current pushes this insulation right up to its rating and no further?

That "30°C, inside a raceway" part is the fine print. It is a laboratory baseline, not a jobsite. The table gives you a starting point that assumes the heat has somewhere to go and the surroundings are cool. The real world rarely cooperates.

Two thieves: ambient heat and crowded conduit

Two conditions steal ampacity, and the Code corrects for each separately.

The first is ambient temperature. If the air around the conduit is already warm, the insulation has less thermal headroom before it hits its limit. A conduit baking in an attic, a run along a sunlit exterior wall, or a pull through a boiler room is not living at 30°C. Table 310.15(B)(1) lists correction factors for this. At a 40°C ambient, a 90°C conductor keeps only about 91% of its rated ampacity. In a genuinely hot space — say 50°C — that drops to roughly 82%.

The second thief is bundling, what the Code calls adjustment for more than three current-carrying conductors. Heat that one wire sheds easily becomes a problem when nine wires share a pipe, each one warming its neighbors and none of them able to dump heat efficiently. Section 310.15(C)(1) handles this with a sliding scale: four to six current-carrying conductors keep 80% of their ampacity, seven to nine keep 70%, ten to twenty keep 50%. The crowd doesn't make any single wire carry more current — it makes the same current run hotter.

The two factors stack. Multiply them both against the base number, and a generous table value can shrink quickly.

A worked example

Take that #6 copper THHN. In the 90°C column of Table 310.16, it shows 75 amps.

Now put it on a real job: six current-carrying conductors sharing a conduit, run through a space that reaches 40°C. Apply the ambient correction (0.91 for 40°C at the 90°C column) and the bundling adjustment (0.80 for four-to-six conductors):

75 × 0.91 × 0.80 ≈ 54.6 amps.

The wire that looked good for 75 is good for about 55 under those conditions. If you'd sized it to a 60-amp continuous-ish load on the strength of that table number, you would have built in a quiet overheat from day one — no nuisance trip, no obvious fault, just insulation living above its design temperature year after year.

Why you start at the 90°C column but can't end there

Here is the part that trips up even experienced hands. You ran the derating math off the 90°C column — 75 amps — even though #6 in the 75°C column is only 65 amps. Is that legal?

Yes, and it's deliberate. The Code lets you use the conductor's full 90°C ampacity as the starting value for correction and adjustment, because those factors are about the heat the wire generates internally. A 90°C insulation genuinely tolerates more heat before failing, so it has more to give back after derating.

But there's a second, independent limit that has nothing to do with bundling or ambient: the terminations. Section 110.14(C) ties you to the temperature rating of the lugs, breakers, and devices the wire lands on. Most modern equipment terminals are rated 75°C. The screw on a breaker doesn't care that your insulation survives 90°C — it was tested and listed at 75. So your final ampacity can never exceed the 75°C column value, regardless of how the derating math comes out.

The rule, stated plainly: derate from the 90°C column, then compare the result against the 75°C termination limit, and use whichever is lower. For our #6, the derated 90°C figure (54.6 A) sits below the 75°C termination cap (65 A), so 54.6 A governs. Flip the conditions — fewer conductors, a cool basement — and the derated 90°C number might climb above 65, at which point the termination rating becomes the ceiling instead. You always have to check both gates.

Where this bites people on real jobs

The classic failure is the full conduit. Someone runs three separate 20-amp circuits — six current-carrying conductors plus a shared neutral situation — through one pipe to save a pull, never applies the 80% adjustment, and the homeowner reports flickering and warmth months later. Nothing is faulted. Everything is just running hotter than it should.

Rooftop and exterior runs are the other repeat offender. A conduit in direct sun can sit well above the air temperature in the shade beside it, and that elevated ambient is exactly what the correction factors exist to catch. The wire is fine in March and marginal in July, which is precisely the kind of intermittent, weather-linked symptom that makes a callback so hard to diagnose.

Neither of these is a knowledge gap so much as a step-skipping gap. The electrician knows the factors exist. The table number is just sitting right there, already looking like an answer, and under deadline it's easy to treat the ceiling as the promise.

Doing the full calculation, every time

The honest version of conductor sizing is three steps, not one: read the base ampacity from the right insulation column, multiply by the ambient correction and the bundling adjustment, then confirm the result doesn't exceed the termination rating. Skip any of the three and you're sizing to a condition that doesn't exist on the job in front of you.

This is exactly the arithmetic Voltly is built to keep honest. Punch in the conductor, the ambient temperature, and how many current-carrying conductors share the raceway, and it applies the correction and adjustment factors from the current tables, then holds the result against the 75°C termination limit — the same three gates, in the same order, without the temptation to stop at the first number that looks finished. It runs entirely offline, which matters when the conduit you're worried about is in a mechanical room with no signal.

If you've ever sized a wire off the table and then wondered, halfway home, whether you remembered the bundle, you can check the whole calculation in the truck before you pull — voltly.lumenlabs.works.