What Ambient and Bundling Do to a Conductor

Why this matters

Ambient temperature and conductor count both reduce ampacity, they get applied as a pair of multipliers, and that is where most techs stop thinking about them. They are not the same kind of thing. One of them is a closed-form relationship you can derive in your head and check against the table. The other is a stepwise allowance for a geometry nobody measured, which is exactly why it cannot be interpolated, extrapolated or argued with. Treating them as interchangeable produces two field errors that look opposite and are the same mistake: interpolating a bundling step to squeeze a size, and refusing an ambient correction on a run that only gets hot two weeks a year.

And there is a third question sitting underneath both, which is which conductors in the pipe count. That one is a judgment about the loads, not a count of wires, and the same raceway with the same wires in it can land on two different answers.

Two terms in the same balance

A conductor's ampacity is the current at which its own heating holds the insulation at its temperature limit: heat in as current squared times resistance, heat out at a rate set by the temperature difference divided by the thermal resistance of the escape path. (The conditional-rating card owns the general form; this one takes the two terms apart.)

Ambient temperature attacks the temperature DIFFERENCE. Raise the surrounding air and you have removed part of the head of pressure that pushes heat out, without touching the path.

Bundling attacks the PATH. More conductors in the raceway means a longer, more obstructed route from the middle of the bundle to the outside air, and it means every neighbour is also a heat source, so the conductor is not just insulated better, it is being warmed by its company.

       room air at ambient temperature
   +---------------------------------------+
   |  raceway wall                         |
   |     o      o      o                   |
   |     o    center   o                   |
   |     o      o      o                   |
   +---------------------------------------+
   heat leaving the center conductor crosses
   its neighbours first, and each neighbour
   is itself a heat source, so the path is
   both longer and hotter than one conductor
   in the same raceway would have seen

Because the two act on different terms, they MULTIPLY. They do not average and they do not compete. Whoever wrote each correction assumed the other would also be applied.

Ambient, which you can derive

Hold the thermal path constant and the insulation limit constant, and the current allowed scales with the square root of the temperature difference available. Heat generated goes as current squared, heat removed goes linearly with the temperature difference, so setting them equal makes current proportional to the square root of that difference.

That gives the correction factor as the square root of (the conductor's temperature rating minus the new ambient) divided by (the conductor's temperature rating minus the reference ambient), where the standard ampacity table's reference is 30 degrees C. The adopted NEC publishes both the correction table and the relationship it comes from in Article 310, so this is a check on your table reading rather than a substitute for it.

Run it against a value from the table. For a 90 degrees C conductor at a 45 degrees C ambient: 90 minus 45 is 45, 90 minus 30 is 60, 45 divided by 60 is 0.75, and the square root of 0.75 is 0.866. The published factor for that row is 0.87. It closes.

Two conditions travel with that relationship. It is derived assuming the thermal path is UNCHANGED, which is why it says nothing about bundling, and assuming conductor resistance is constant, when copper's resistance rises with temperature, so the true relationship is slightly less favourable than the square root suggests. Use the formula to check your table reading, not to replace it.

Check the direction at the other end of the range, because this is where techs get it backwards. Below the 30 degrees C reference the factor goes ABOVE 1.0: a run in a cold space is permitted more current, not less, and the adopted code's table says so. If your understanding of ambient correction cannot produce a factor greater than one, it is a memorised penalty rather than a relationship.

Bundling, which you cannot

There is no equivalent formula, and the reason is worth knowing. The real answer depends on where in the bundle a conductor sits, how tightly it is packed, whether the raceway is metallic, and whether the circuits are loaded at once or take turns. None of that is knowable from a count, so the adjustment is a conservative step function covering the range of geometries a count is consistent with.

In the adopted NEC, the adjustment for more than three current-carrying conductors in a raceway or cable runs 80 percent for four through six, 70 percent for seven through nine, and 50 percent for ten through twenty, with further steps above that.

Two consequences fall straight out of it being a step table. Do not interpolate. Seven conductors and nine conductors both get 70 percent, and that is the table working as designed rather than a rounding error you are entitled to fix. And the steps are large, so one added conductor at the wrong place in the count costs 10 percentage points at once. Going from six to seven conductors takes 80 percent to 70 percent, which is a 12.5 percent cut in the corrected ampacity for a single wire.

There are conditions where the adjustment does not apply at all, including short nipples and certain cable and tray arrangements, and those live in the same Article of the adopted code. Read them there rather than carrying a remembered exception, because they are narrow.

Which conductors count

The adjustment counts CURRENT-CARRYING conductors, and that is a decision about the loads rather than a count of what is in the pipe.

• Equipment grounding conductors and bonding jumpers do not count, carrying no current in normal operation.
• The neutral of a three-phase, four-wire wye system carrying only balanced linear loads does not count, because the phase currents largely cancel in it.
• The same neutral does count where the major portion of the load is nonlinear, because the third-harmonic components do not cancel and instead add in the neutral. The neutral card owns the mechanism and the arithmetic; for this card the point is that a neutral changes category based on what is plugged in downstream.
• The neutral of a three-wire circuit derived from a four-wire wye counts, because it carries roughly what the phase conductors carry.
• Conductors that cannot be energized at the same time as each other may be excluded, which is what lets some control and interlocked arrangements out of the count.

The provisions are in Article 310 of the NEC edition your authority having jurisdiction has adopted, and they are the place to read the exact wording rather than a summary.

Worked case: one raceway, two answers

A raceway carries two three-phase, four-wire branch circuits and one equipment grounding conductor: six phase conductors, two neutrals, one grounding conductor, nine wires total. The conductors are 6 AWG copper with 90 degrees C insulation, base ampacity 75 A in the 90 degrees C column, landing on terminals listed for 75 degrees C, where 6 AWG copper is 65 A. The run passes through a space that reaches 40 degrees C in summer. The load on each circuit is 40 A continuous, so the required ampacity is 40 times 1.25, or 50 A.

Ambient first, applied to the 90 degrees C base. The factor is the square root of (90 minus 40) over (90 minus 30), which is the square root of 50 over 60, or the square root of 0.8333, which is 0.913. Call it 0.91.

Case one: the loads are three-phase motors. Balanced, linear, so neither neutral counts and the grounding conductor never counts. Current-carrying conductors: six. The adjustment is 80 percent.

Corrected ampacity is 75 times 0.91 times 0.80, which is 54.6 A. The termination cap of 65 A is higher, so it does not bind. Against the required 50 A, the circuit passes with about 9 percent margin.

Case two: same raceway, single-phase electronic equipment fed line to neutral. The major portion of the load is nonlinear, so both neutrals count. Current-carrying conductors: eight. And the count is not the whole of it. Line-to-neutral nonlinear load does not merely make each neutral countable, it makes it the most heavily loaded conductor in the raceway, because triplen components add arithmetically instead of cancelling. At third-harmonic content near 60 percent of the fundamental each neutral carries roughly 1.5 times its phase conductors, about 60 A here against a corrected ampacity of 47.8. The phase shortfall is the smaller of this raceway's two problems, and the neutral is the one with no overcurrent device watching it. The adjustment steps down to 70 percent.

Corrected ampacity is 75 times 0.91 times 0.70, which is 47.8 A. The termination cap is still 65 A and still does not bind. Against the required 50 A, the circuit fails, short by 2.2 A.

Identical raceway, identical wire, identical ambient, identical load current. The corrected ampacity moved from 54.6 A to 47.8 A, which is 12.5 percent lower, and the verdict flipped. Nothing physical changed. What changed was an answer to a question about the loads, and that answer is not visible from the outside of the conduit.

The failure mode this creates

The dangerous version of case two is not a new install, where somebody at least had to think about it. It is a retrofit: a space full of linear single-phase loads gets replaced with electronic ones during an upgrade, the electrical work is a like-for-like reconnection, and nobody re-runs the adjustment because no conductors moved. The two neutrals silently changed category, and the raceway lost a step of adjustment on the same afternoon its neutral current went up. It shows up as a raceway running warmer than its neighbours, insulation aging years early, and eventually a fault at the point with the worst escape path, which is usually a long horizontal section above a ceiling rather than the panel where anyone was looking.

How to verify you got this right

Three field checks, and none of them are recomputations of the table.

Measure the ambient where the run is hottest, in the season it is hottest. Not at the panel, and not in February. A run through a ceiling space above a mechanical room, or a rooftop section in the sun, sees an ambient the panel never does, and it is the worst section of the run that sets the correction. If reaching the run means a ladder or a lift, set the ladder on a level surface at the correct angle and rated for you plus your tools, or use a lift; do not climb on the raceway or the equipment.

Count the conductors in the raceway, de-energized, and count the neutrals separately. 29 CFR 1910.333(a)(1) requires the circuit de-energized before work on or near exposed live parts unless the employer can demonstrate that de-energizing introduces additional or increased hazards or is infeasible due to equipment design or operational limitations, and a conductor count can always wait. Lock and tag under 29 CFR 1910.333(b)(2), which is the standard that governs electrical work since 1910.147 excludes exposure to electrical hazards from work on conductors and equipment in electric utilization installations at (a)(1)(ii)(C), with 29 CFR 1926.417 as the construction counterpart, and prove dead live-dead-live per NFPA 70E-2021 120.5, in the edition your employer's electrical safety program or your authority having jurisdiction has adopted.

Settle the nonlinear question with a measurement rather than an opinion. Clamp each neutral under load, an energized reading taken under the same gate with PPE selected at NFPA 70E-2021 130.5 and 130.7. A neutral reading a small fraction of the phase currents behaves as the balanced-linear case describes; one reading a substantial fraction of them, or more, is a current-carrying conductor by behaviour and the count has to include it. That single reading settles the difference between the two cases above faster than arguing about the load inventory.

References

• NEC Article 310, ambient temperature correction, adjustment for more than three current-carrying conductors, and the treatment of neutral conductors, in the edition your authority having jurisdiction has adopted
• 29 CFR 1910.333(a)(1) and 1910.333(b)(2), with the 1910.147(a)(1)(ii)(C) exclusion; 29 CFR 1926.417 for construction
• NFPA 70E-2021, 120.5, 130.5 and 130.7, as adopted by your employer's electrical safety program or your authority having jurisdiction
• See related: What Derating Is, and Why a Conductor Rating Is Conditional; The Derating Table and When It Applies to You; Why a Neutral Can Carry More Than You Expect