How Heat and Current Relate in a Conductor
Why this matters
Two identical conductors, carrying the identical current, sitting in two buildings: one runs warm for thirty years and the other bakes its insulation brittle in three summers. Nothing about the wire is different. What differs is the thermal environment it lives in, and that is the part techs skip when they size or judge a run by looking only at the gauge and the load. Ampacity is not a property of the conductor. It is a property of the installation, and reading it that way turns a table lookup into an actual diagnosis of why a run is failing.
Before you touch a run you suspect is hot
Do not judge a conductor's temperature by hand. A conductor at the top of its insulation rating will burn skin, and the one you reach for may be the one that is degraded. Use a non-contact instrument.
Opening a panel door to take a thermal image is energized work, permitted under the troubleshooting exception at 29 CFR 1910.333(a)(1) only where the assessment cannot be made with the equipment de-energized, and it requires the arc-flash risk assessment at NFPA 70E-2021, 130.5 with the protection that assessment calls for. An infrared window in the enclosure lets you take the same image with the door closed and is the better answer wherever one exists. Any instrument you connect must be rated for the system under 29 CFR 1910.334(c)(2).
Work on the conductors themselves is done de-energized, locked and tagged under 29 CFR 1910.333(b)(2), or 29 CFR 1926.417 on construction work, and proved dead with the live-dead-live sequence at NFPA 70E-2021, 120.5. One practical rule about what you find: insulation that has been run hot for a long time goes hard and cracks when it is moved, so a run confirmed overheated is replaced rather than re-terminated and re-used. Pulling on brittle insulation strips it and leaves bare conductor inside a raceway you cannot see into.
Heat in, heat out
A conductor's temperature is not set by how much heat it makes. It is set by the balance between how much it makes and how fast its surroundings take that heat away. Both halves have to be in the picture or the answer is wrong.
Generation is current squared times resistance. The squared term is the one that surprises people. Raise a load from 24 A to 30 A, a 25 percent increase, and the heat generated goes up by 30 divided by 24, squared, which is 1.56 - a 56 percent increase in heat for a 25 percent increase in current. Halve the current and you cut the heat to a quarter. This is why a modest, chronic overload is a serious thermal event and why a brief overload usually is not.
Resistance is the other factor, and it is not fixed. Copper's resistance rises with temperature by roughly 0.4 percent per degree C, so a conductor that heats up generates more heat at the same current than it did when cool. That is a mild positive feedback in a healthy run and a runaway at a bad connection, which a sibling article on why a loose connection gets hot covers in its own right.
Removal is the temperature difference between the conductor and its surroundings, divided by the thermal resistance of the path out: through the insulation, across any air gap, through the raceway wall, into the surrounding air. Two things wreck that path. Raising the ambient temperature shrinks the difference that drives the heat out. Adding more current-carrying conductors in the same raceway adds heat sources and simultaneously blocks each other's escape route.
What the copper reaches is where those two curves cross. The limit is not the copper, which is indifferent to anything you will meet in a building. The limit is the insulation's temperature rating, and beyond it the insulation stops being insulation.
The gate
A conductor is thermally acceptable for a given load when both of the following are true, joined by AND:
- Its ampacity, taken from the table for its size and insulation rating and then multiplied by the ambient correction factor and the conductor-count adjustment factor for its actual installation, is at or above the required current, where a continuous load - one whose maximum current is expected to continue for three hours or more - requires 125 percent of the load current.
- The final value does not exceed the ampacity for the temperature rating of the terminations at both ends, because a lug rated 75 degrees C limits the circuit regardless of what the conductor's insulation is rated for.
The figures below are used to show the method. Read the ampacity, correction and adjustment tables in the edition of the electrical code your jurisdiction has adopted, because both the values and the section numbering have moved between editions.
Same wire, same load, two verdicts
A 24 A continuous load on a 10 AWG copper conductor with 90 degree C insulation, terminations rated 75 degrees C at both ends. Required current is 24 times 1.25, which is 30 A.
Installation A. Three current-carrying conductors in a raceway, ambient 30 degrees C, which is the temperature the base table is written for. No ambient correction, and the conductor-count adjustment does not begin until there are more than three current-carrying conductors, so no adjustment either. The 75 degree C column value for this conductor is 35 A. Required 30 A against 35 A available: the first half of the gate passes. Terminations are rated 75 degrees C and the 75 degree C column is what we used, so the second half passes. This run is acceptable, and it will live a long life.
Installation B. Same conductor, same 24 A continuous load, but six current-carrying conductors share the raceway and the room runs at 45 degrees C.
Start from the highest column the conductor's insulation permits, which is the 90 degree C column at 40 A, because the correction and adjustment factors are applied there. The adjustment factor for four to six current-carrying conductors is 80 percent. The ambient correction factor in the 90 degree C column for 45 degrees C is 0.87. So 40 times 0.80 times 0.87 gives 27.8 A.
Required is still 30 A. Available is 27.8 A. The first half of the gate fails, by about 2.2 A, roughly 7 percent short. Check the second half anyway, because it teaches which constraint is actually binding: the 75 degree C termination column value is 35 A, and 27.8 A is below it, so terminations are not the limit here. The limit is entirely thermal environment.
Nothing about the wire changed between A and B. The load did not change. The two answers differ because six conductors in a hot room cannot get rid of the heat that three conductors in a 30 degree C room shed easily, and the code's factors are simply the arithmetic version of that sentence.
What the correction factors are actually saying
Techs treat these tables as bureaucracy. They are a thermal model with the physics already done:
- The ambient correction factor is the shrinking temperature difference. A 90 degree C conductor in a 30 degree C room has 60 degrees of margin to push heat across. In a 45 degree C room it has 45 degrees, three quarters as much, and the factor reflects that.
- The conductor-count adjustment factor is mutual heating. Each conductor is a heater; each one is also insulation for its neighbours. The factor steps down as the count rises because both effects compound.
- The termination limitation is a different mechanism entirely, and it catches people who correctly did the first two. A lug is a thermal bridge into a device, and the device's own materials have a rating. Landing a 90 degree C conductor on a 75 degree C lug does not make the lug 90 degree C rated; it makes the circuit 75 degree C limited.
What flips the answer
Three conditions genuinely change the method rather than just moving a number.
Free air instead of a raceway. A conductor in free air sheds heat far better and carries a substantially higher ampacity from a different table. Applying the raceway table to a free-air run is conservative and wasteful; applying the free-air table to a raceway run is dangerous.
A short high-ambient segment. A run that passes briefly through a hot zone is not the same as a run that lives in one, and the code carries provisions for short lengths that a blanket correction would ignore. Check the actual adopted text rather than assuming either way.
A non-continuous load. The 125 percent multiplier exists because a conductor takes time to reach its final temperature and a three-hour load has all the time it needs. A load that peaks briefly and falls away does not reach thermal equilibrium at its peak, which is why intermittent duty is treated differently.
Finding it before the insulation does
The instrument that catches this is a thermal image taken under real load, and the discipline point is that the image shows you the outside of things. A raceway's exterior is cooler than the conductors inside it, and a conductor buried in a bundle is hotter than the one on the outside of the bundle that you can see. So a thermal image of a conduit or a full gutter is a lower bound, not a measurement.
Take images at real load, not after a service call has had the equipment off for an hour, and record the load current alongside the image the way you would with any thermal reading. An image with no current beside it cannot be compared to next year's image. Where the image shows a warm raceway with no single hot spot, you are looking at a distributed loading and environment problem of the kind Installation B has. Where it shows one bright point on an otherwise cool run, you are looking at a connection, which is a different fault with a different fix.
How to verify you got this right
Recount the current-carrying conductors physically rather than from the drawing, and measure the ambient with an instrument in the actual space at the hottest part of the day the equipment runs. Both of those inputs are routinely assumed and both routinely turn out wrong: a raceway that was drawn with three conductors has five in it, and a mechanical room that is nominally conditioned runs well above what anyone claimed.
Then re-run the arithmetic with your measured inputs and confirm which half of the gate binds. If your corrected ampacity now clears the required current with margin, and the terminations are rated for the column you used, the run is right. If it clears by only a percent or two, treat that as failing, because your ambient measurement is a snapshot and the hottest day of the year is not the day you measured.
References
- NFPA 70 (National Electrical Code), Article 310 for conductor ampacity and the ambient-correction and conductor-count adjustment factors, and Article 110 for termination temperature limitations - use the edition adopted in your jurisdiction
- 29 CFR 1910.333(a)(1) and (b)(2) - the troubleshooting exception and safe work practices for electrical work; 29 CFR 1926.417 for the construction counterpart
- 29 CFR 1910.334(c)(2) - test instruments and equipment rated for the circuits to which they are connected
- NFPA 70E-2021, 120.5 (establishing an electrically safe work condition) and 130.5 (arc flash risk assessment)
- See related: Why a Loose Connection Gets Hot; What a Hot Spot in a Thermal Image Actually Means; Conduit Fill and Wire Ampacity Reference