Conductor ampacity — SELLC Engineering Calculators

Result

Allowable ampacity —

Governing limit —

Conductor —

Columns, wire / term —

Ambient used —

The working

The ladder below is the product of this page. A bare ampere figure would be an oracle; what an engineer needs is the arithmetic, the table each factor came out of, and the point at which one limit overtook the other. Read it top to bottom — the order is not interchangeable, and step 5 is deliberately not corrected.

Step	What	Basis — NEC	Value

Base, Table 310.16 —

Corrected and adjusted —

Termination column, uncorrected —

Count and adjustment —

Preliminary sizing only. Not for construction. This page is a teaching and checking aid. It is not a substitute for a licensed engineer’s sealed design, it does not know your installation, and it does not relieve anyone of the obligation to work from the printed Code. Values are from the current NEC edition as adopted; your authority having jurisdiction may have adopted a different edition or amended it. Verify every figure against the printed table before it leaves your desk.

What this page does not do

These are permanent boundaries, not caveats on one answer. They do not go away when the inputs change.

Controls

Conductor

Material

Table 310.16 carries copper and aluminum (or copper-clad aluminum). Aluminum has no 14 AWG row; switching material rebuilds the size list from the table itself.

Size

Built from the table, never typed — every size offered is a row that exists, so a size off either end of Table 310.16 is not reachable.

Conductor insulation rating

The conductor’s own rating. This is the column steps 1 and 2 are read from, and it is the one people get wrong.

Termination

Equipment termination rating

What the lugs, breaker jaws and terminal blocks at both ends are listed for — a property of the equipment, not of the wire. 110.14(C) caps the answer at the uncorrected Table 310.16 value in this column.

Circuit amperes 200 A

Read for two things, neither of which is sizing. First, the 110.14(C)(1) boundary: paragraph (a) is “100 amperes or less” and 100 A is inside it; paragraph (b) is “over 100 amperes.” Second, a caveat beside the result when the current you enter is above the ampacity computed here — a remark about two numbers already on this page, not a size check. No continuous-load multiplier is applied to it and no overcurrent device is chosen from it.

Installation conditions

Ambient temperature 30 °C

The ambient the raceway or cable sits in — not the room next door, and not the annual average. The slider deliberately runs past both ends of the correction data this page carries, so you can see where it refuses.

Raceway or cable in sunlight on or above a roof

310.15(C)(2) moves the ambient before the correction table is opened, rather than scaling anything after it. This is the single most common way a public ampacity answer comes out wrong, because the reader types the weather. The adder is conditional in two ways this tick cannot check for you: it is written against a clearance above the roof surface, and there is an exception for Type XHHW-2. This page checks neither and applies the adder whenever the box is ticked. See The rooftop adder below.

Current-carrying conductors 3

In the same raceway or cable, bundled together. Count what actually carries current — see the 310.15(E) caveat with every result. Past the top of Table 310.15(C)(1) the open “41 and above” band applies; it is a printed row, so it is used, not refused.

Circuit voltage

Selects which rules apply — Table 310.16 by its own heading, and 110.14(C) at 1000 V and under versus 110.40 above it. Not used for voltage drop; no voltage-drop calculation is performed on this page.

The whole table, at your conditions

Ampacity against conductor area

Base, Table 310.16 at the conductor’s column (fine dotted line)
Termination column, uncorrected — the 110.14(C) cap where that section applies (long dashed line)
Corrected and adjusted (short dashed line)
Allowable ampacity (solid line, drawn thicker)

Every row of Table 310.16 at the current settings.

The horizontal axis is conductor area in thousands of circular mils, on a logarithmic scale, because that is the axis on which ampacity is nearly a straight line — the table is not linear in area, and reading it on a linear axis crushes everything below 4/0 into the first few pixels.

What the two dashed curves usually show is that there is no crossing. The corrected and adjusted curve is the base curve scaled by one factor that does not depend on the conductor size, and the termination curve is simply another column of the same table — so their ratio is nearly constant from one end of the table to the other, and in the great majority of settings one of them is the lower one at every row. That includes the settings this page loads with, where the termination cap governs from the smallest row to the largest and the solid line lies on the dashed cap curve across the entire width. The reading to take is therefore not a crossing point but which dashed curve the solid line is sitting on, and over how much of the table.

Above 1000 V the termination curve stops being a limit at all — 110.40, not 110.14(C), is the section written for the terminations there, and it is a listing condition rather than a table lookup, so it yields no ampere value to compare against. Set the voltage control to 2000 V and the solid line stops being clipped by the cap curve: it follows the corrected and adjusted curve at every row, which on a warm ambient or a full conduit leaves it well below the termination curve and on a light one leaves it above. Above is not a drawing error and it is not an ampacity the cap is being ignored for — on that circuit the dashed termination curve is a reference line, and step 5 of the ladder labels it as one.

Where the two curves do change places, it usually runs the opposite way to the intuition. The ratio between any two columns of Table 310.16 is not constant: it is at its widest at the smallest sizes and generally narrower at the large end, though it does not narrow step by step. So when the correction and adjustment factors happen to land the two curves within a few per cent of each other, the common pattern is the termination cap binding at the small end and the conductor’s own limit taking over as the conductor grows. Read that as the usual shape and not as a rule: there are settings — aluminium on the 75 °C column into 60 °C terminations, at the cold end of the ambient range, is one — where the cap governs at both ends of the table and the conductor’s own limit takes over only at scattered sizes in between. Because that column ratio is not monotonic row by row, the two can swap back and forth a dozen times across the width of the table rather than cross once cleanly, and which curve is on top at any one size is a thing to read off the ladder rather than to predict. Move the ambient or the conductor count and the corrected curve slides bodily up or down against a fixed cap curve, which is the whole argument for drawing both rather than printing one answer.

The method, and where it stops

The order of operations

Step 1 — base ampacity, Table 310.16

I_base = Table 310.16 [material, size, conductor column]

Step 2 — ambient correction, Table 310.15(B)(1)

F_amb = Table 310.15(B)(1) [ambient, same column as step 1]

Step 3 — adjustment for conductor count, Table 310.15(C)(1)

F_adj = Table 310.15(C)(1) [number of current-carrying conductors]

Step 4 — the corrected and adjusted ampacity

I_ca = I_base · F_amb · F_adj

Step 5 — the termination limit, 110.14(C)

I_term = Table 310.16 [material, size, termination column], uncorrected

Step 6 — the answer, where 110.14(C) is the rule for the terminations

I_allow = min( I_ca , I_term )

Step 6 — the answer, on circuits above 1000 V

I_allow = I_ca

I_base — the table value for not more than three current-carrying conductors in a raceway, cable or earth, at the table’s own reference ambient. Everything after this is a departure from those two assumptions.
F_amb — greater than 1 below Table 310.15(B)(1)’s unity band, exactly 1.00 inside it, and less than 1 above it. The unity band ends at the table’s reference ambient rather than beginning there, so a design ambient a few degrees under the reference buys nothing at all while the very first degree above it already costs — step 2 of the ladder prints the band it actually read, and that, not a boundary quoted in prose, is the thing to check. It is a credit or a penalty against the hottest ambient the run will see, not the average one.
F_adj — a mutual-heating penalty, not an electrical one. Bundled conductors heat each other and share one thermal path out.
I_term — a property of the equipment, read off the conductor table only because that is the yardstick 110.14(C) chose. On circuits above 1000 V the section written for the terminations is 110.40 rather than 110.14(C), and 110.40 is a listing condition rather than a table lookup: it produces no competing ampere value, so there is nothing for step 6 to take the lesser of and the answer is I_ca alone. The calculator still shows the termination column there, labelled for reference, because it orients a reader who knows the 110.14(C) arithmetic — but it is not a limit on that circuit, and step 5 of the ladder says so in as many words.

Why step 2 uses the conductor’s own column. This is the error worth naming, because it is a common one in production sizing software: correcting the ambient on the termination column instead. Do that and 90 °C insulation buys you nothing at all, which is plainly wrong — the whole commercial reason to specify THHN/THWN-2 over THW is that the derating arithmetic runs on the 90 °C column even when the final answer may not exceed the 75 °C value. Physically, steps 2 and 3 are about how hot the copper gets in your conduit, and how hot the copper may get is set by the insulation wrapped around it. The termination has no say in that; it gets its say in step 5.

Why step 5 is not corrected. 110.14(C) is not a thermal calculation about the conductor. It is a statement about the lug, the breaker jaw and the polymer around them — parts that sit inside the equipment enclosure, in the equipment’s own ambient, rated by the equipment’s listing and tested by the listing laboratory at that rating. Nothing about how many conductors share the reader’s conduit, and nothing about how hot the roof gets, changes what the terminal was listed for. So no correction and no adjustment is applied to it. Applying them would be double-counting a derate the terminal never earned.

A consequence worth internalising: the two limits are not commutative and they are not interchangeable. It is entirely normal for the ambient correction to raise the table value and for the answer not to move at all, because the termination cap was already the binding limit. This page says so explicitly when it happens rather than leaving the reader to wonder why a factor greater than 1 changed nothing.

The three columns, and the 100 A boundary

Table 310.16 has three temperature columns because insulation systems have three common continuous-operating ratings, and the ampacity in each column is the current that drives the conductor to that temperature under the table’s reference conditions. The columns are not a menu. The one you may read step 1 from is set by the insulation actually installed; the one that caps the answer is set by the equipment at each end.

Column	Typical types	Where it usually binds
60 °C	TW, UF	The 110.14(C)(1)(a) default at 100 A and below
75 °C	RHW, THW, THWN, XHHW, USE	The 110.14(C)(1)(b) default over 100 A, and what nearly all listed equipment carries
90 °C	THHN, THHW, THWN-2, XHHW-2, RHW-2, USE-2	Correction and adjustment arithmetic only — very rarely a termination rating

110.14(C)(1)(a) is “100 amperes or less” and it is inclusive. A 100 A circuit is inside paragraph (a), not paragraph (b) — a boundary it is easy to slip by one, because “over 100 amperes” reads at a glance as “100 amperes and up.” Paragraph (a) defaults to the 60 °C column; paragraph (b), “over 100 amperes,” defaults to 75 °C. Both defaults are displaced only where the equipment is listed and identified for a higher temperature, which is a label you have to go and read. Marking a panelboard “75 °C” on a drawing does not make it so.

Counting current-carrying conductors

Step 3 is the step a user is most likely to feed a wrong number, because the count is a judgement about the circuit and the calculator can only take what it is given. 310.15(E)(1): a neutral carrying only the unbalanced current of the other conductors of the same circuit is not counted. 310.15(E)(2): the neutral of a 3-wire circuit taken from a 4-wire 3-phase wye system is counted. 310.15(E)(3): on a 4-wire 3-phase wye circuit whose load is majority non-linear — LED drivers, VFDs, switch-mode supplies, most modern lighting — the neutral is counted, because the triplen harmonics add in it instead of cancelling and it can carry more current than any phase. 310.15(F): grounding and bonding conductors are never counted. The realistic failure is entering 3 for a shared-neutral wye branch feeding electronic ballasts, which under-counts by one and lands the run in the wrong adjustment band.

The rooftop adder

310.15(C)(2) is filed under adjustment factors but behaves as an ambient change: where a raceway or cable is exposed to direct sunlight on or above a rooftop, 33 °C is added to the outdoor ambient before Table 310.15(B)(1) is opened. It is not a multiplier and it does not commute with anything — a 35 °C design ambient becomes 68 °C, which is beyond the correction data this page carries and produces a refusal rather than a number. A great many undersized rooftop PV and HVAC feeders exist because the designer corrected for the weather station instead.

Check the height before you tick the box. The rule is written against a clearance: it reaches a raceway or cable less than 7/8 in. (22 mm) above the roof surface, and a run carried clear of the roof on standoffs at or above that clearance is outside it and takes no adder at all. Ticking the box for such a run does not merely overstate the ambient by 33 °C — because the adder can push the ambient past the top of the range this page carries, it can turn an installation the Code answers into a refusal, which is its own kind of wrong answer. There is also an exception for Type XHHW-2; read it in the edition you are enforcing rather than take it from this page, and note that where it is in force it turns on the whole raceway — one conductor of another type puts the adder back on the lot. This calculator applies neither condition. It adds the 33 °C whenever the box is ticked: XHHW-2 or not, at any height.

Where this calculator refuses, and whose limit that is. The correction data here covers a bounded ambient range, and outside it this page prints nothing. That is a limit of this data copy, not of the Code: the printed Table 310.15(B)(1) continues above the top of this range and has an open band below the bottom of it. Refusing where the Code answers would be its own kind of dishonesty, so the refusal says which limit it is hitting. It refuses downward for a second reason — the cold-end factors are greater than 1, and quietly extrapolating them would hand out ampacity the table never granted. The one place this page does not refuse is the top of Table 310.15(C)(1): “41 and above” is a printed row, so it is applied.

What to try

Watch 90 °C insulation earn its keep, then hit the ceiling. Copper, 4/0 AWG, 75 °C terminations, 3 conductors, 30 °C. Switch the conductor column between 75 °C and 90 °C: the answer does not move — but not for the same reason in the two cases, and the ladder says so. On the 75 °C column the corrected and adjusted value and the termination cap land on the same number, so neither has slack: the governing card reads “Both, equally” and two rows of the ladder are tagged. On the 90 °C column the conductor’s own value is the higher one and the cap alone governs. Now raise the ambient to 41 °C. Both columns fall below the cap, so the cap stops binding altogether, and the 90 °C conductor keeps a materially higher number than the 75 °C one from the same wire in the same conduit — that difference is exactly what step 2 reading the conductor’s own column buys, and it is what an implementation that corrects on the termination column destroys.
Extinguish an ambient credit. Drag the ambient down until step 2 of the ladder shows a factor above 1.00 — the ladder names the band it read, so you can see exactly where the unity band ends rather than taking a boundary on trust. Step 4 then rises above the base value, and the answer stays put, capped by step 5. The page says so in as many words rather than letting you conclude the multiplication did something. Then go and find the other half of that rule: copper 700 kcmil, a 75 °C conductor into 60 °C terminations, 4 conductors, ambient at the bottom of its range. The caveat changes wording, because there the credit is throttled rather than discarded — strike it out and the answer really would be lower, so the cooler ambient did buy amperes, up to the cap and no further. “The cap governs” and “the credit bought nothing” are different claims, and this page prints the second only where it is true.
Cross the 100 A boundary. Leave the terminations at 75 °C and slide circuit amperes from 100 to 110. At 100 A the page warns that 110.14(C)(1)(a) defaults to the 60 °C column and that 100 is inside that band; at 110 A the warning goes away, because paragraph (b) defaults to 75 °C.
Trigger the rooftop refusal. Ambient 35 °C, then tick the rooftop box. 35 + 33 = 68 °C, past the top of the correction data, and the result panel is replaced by a refusal that names the arithmetic. Drop the ambient to 15 °C and tick it again: 48 °C computes normally, and the ladder shows the adder explicitly rather than hiding it inside a factor.
Fill a conduit, and watch a whole curve drop in one step. Slide the conductor count from 3 to 4. Table 310.15(C)(1) is banded, not continuous, so the adjustment factor goes from 1.00 to 0.80 in a single slider step: the corrected and adjusted curve does not travel across the chart, it jumps, and it lands below the termination curve at every row but the smallest, where the two come out exactly equal. There is no crossing point sliding along the axis to watch — there is a two-frame flip, after which the conductor’s own limit governs across the whole table and the termination cap never binds again. Keep sliding: the curve drops again at 7 and again at 10, each time by a whole band and never in between.

Assumptions and limits

Table 310.16 conditions only. Not more than three current-carrying conductors in a raceway, in a cable, or directly buried, at the table’s reference ambient. Single conductors in free air are Table 310.17 and are substantially higher; cable tray is further modified by 392.80. Neither is computed here and neither may be inferred from this page.
This is an ampacity, not a conductor size. No load is applied, no 125 % continuous-load multiplier from 210.19(A)(1) or 215.2(A)(1), no comparison against an overcurrent device, no voltage drop. The circuit-amperes control resolves the 110.14(C)(1) boundary, and where the current entered is above the ampacity computed here it says so — which is a remark about two numbers on this page, not a sizing step and not a verdict on the circuit.
Parallel conductors are not handled, and the figure printed here is per conductor. Where a circuit is made up of paralleled sets, the ampacity above belongs to one conductor: two 4/0 AWG sets carry two conductors’ worth of it, not one. Every conductor of every set is current-carrying, so the count entered above has to be the total in the raceway — two three-phase sets in one conduit are six current-carrying conductors, not three, and the adjustment factor follows the six. 310.10(G) also sets a floor and a list of conditions: conductors are permitted in parallel only in sizes 1/0 AWG and larger, with limited exceptions, and the paralleled conductors must match in length, material, size, insulation and termination method. None of that is checked here.
The adjustment factor is applied unconditionally. Whenever the count above exceeds three, Table 310.15(C)(1) is applied — this page never asks whether one of 310.15(C)(1)’s exceptions relieves it. The best known is a raceway nipple: conductors in a raceway not exceeding 600 mm (24 in.) in length need no adjustment. Read the exceptions in the edition you are enforcing; where one of them applies, the figure here is conservative rather than wrong.
No overcurrent device size and no grounding conductor size is printed. An ampacity is not a device rating. 240.4(D) in particular is a limit on the device protecting a small conductor, not on the conductor’s ampacity — a 12 AWG copper conductor’s ampacity is what the table says; what 240.4(D) changes is the largest breaker permitted to protect it.
One ambient, one bundle, one run. Real runs pass through several thermal environments; 310.14(A)(2) lets a short hotter section be disregarded under conditions this page does not test — and it is 310.14, not 310.15, since the 2020 restructure moved the selection-of-ampacity rules out of 310.15 and gave 310.15 to correction and adjustment. If your run crosses a boiler room, correct for the boiler room.
The tables are a copy, and copies drift. These values are generated from the same tables the firm’s own engineering software reads, by a generator that asserts row counts, contiguity, monotonicity, membership and cross-table agreement; a drift check then compares the generated file byte for byte against a fresh generation. Those are machine checks and they catch drift — a cell that moves after generation, a shape the generator does not understand, a hand edit of the generated file. None of them can catch a cell that was wrong the first time it was transcribed — a plausible wrong value passes every one of them. Only a person reading printed NFPA 70 against every cell catches that, and this page does not claim anyone has. Read any figure you intend to rely on out of the printed Code, and if it disagrees with this page, please say so.
Edition and adoption are yours to confirm. Values shown are the stated NEC edition. Jurisdictions adopt on their own schedule and amend freely; several amend 310.15 and 110.14 specifically.
Not for construction. Preliminary sizing and checking only. No output of this page is a sealed engineering document, and nothing here substitutes for the judgement of the engineer of record.

Why an engineer cares

Ampacity is the most frequently performed calculation in electrical design and the one most often done from memory, which is why it is also a reliable source of field conflict. The failure mode is rarely a wrong table value; it is a right value taken in the wrong order. Correcting on the termination column, forgetting the rooftop adder, counting a harmonic-loaded neutral as a non-current-carrying conductor, or treating 100 A as “over 100 amperes” each move the answer by one adjustment band or one whole column — enough to change a conductor size, and enough to change it in the unsafe direction. A conductor run at above its allowable ampacity does not fail immediately or visibly; it ages its insulation faster, and the consequence arrives years later as a fault in a wall. Showing the ladder rather than the answer is the point of this page: it is checkable by a reviewer who disagrees with it, which a bare number never is.

On the two-copies problem. The NEC tables behind this page are generated from the same source the firm’s internal tools read, and a drift check compares them byte for byte. That is deliberate: a second, hand-typed copy of a code table disagrees silently, and a stamped drawing must not carry a silent disagreement. So every table value the calculator uses — each ampacity, each ambient correction factor, each conductor-count adjustment factor — is read out of that generated copy as you watch, rather than written into this page’s text. Not everything here is covered by that. A handful of Code constants have no home in the generated data yet and are typed into the calculator’s own code: the rooftop adder, the clearance it is written against, the voltage above which 110.14(C) is no longer the section for the terminations, and the 100 A split inside 110.14(C)(1). The prose on this page names those and other Code boundaries in words as well. None of that is covered by the drift check, so read the prose as signposts and take the numbers themselves off the ladder, which came out of the table.