How Many Cables Fit in a Conduit? It Is a Code Question, Not a Geometry Question

Why the limit is not about fitting

Take the calculator's default and reverse the question. Three 12 AWG conductors occupy 7.49 % of a 3/4 in EMT. Sixteen of them occupy 39.92 %. Even at sixteen, three fifths of the tube is air — and a run stuffed to sixty per cent by hand would still not be full in any packing sense, because circles in a circle cannot exceed about ninety per cent even when arranged perfectly and nobody arranges cable perfectly. So the rule is plainly not a statement about volume. It is a statement about two failure modes.

The first is thermal. A conductor carrying current dissipates I²R as heat, and that heat has to leave through the insulation, through the air gaps, through the conduit wall and into the room. Insulation is rated by temperature — the two Hs in THHN stand for high heat resistance, 90 °C in dry locations — and past that temperature it hardens, cracks and eventually stops being insulation. Put one conductor in a pipe and it radiates freely into a lot of still air. Put nine in and the ones in the middle are surrounded by other heat sources with the outer ones between them and the wall. The fill limit is a crude proxy for keeping enough air in the tube that the middle of the bundle can still shed heat.

The second is mechanical, and it is the one that bites during the job rather than years later. Conductors are pulled through conduit with lubricant and, on long runs, with a winch. Every bend multiplies the tension needed at the far end, and the fuller the pipe the less the bundle can rearrange itself around a bend, so it binds. What gives way first is the insulation, scraped against the edge of a coupling or a bushing, and the damage is invisible from either end of the run. A conductor with a nick in its insulation buried in a wall is a fault waiting for a damp winter. The 40 % figure is as much about being able to install the cable safely as about running it safely.

Which code this implements, and the numbers it takes from it

The tool is a National Electrical Code tool, and its title says NEC so nobody has to guess. Underneath it carries two tables copied from NEC Chapter 9. Table 4 gives the total internal cross-sectional area of each conduit at each trade size, and the tool holds five families: EMT, PVC Schedule 40, PVC Schedule 80, rigid metal and intermediate metal, in trade sizes from 1/2 in to 4 in. Table 5 gives the approximate area of an insulated conductor, and the tool holds the THHN and THWN column from 14 AWG up to 4/0. Both sets were checked against the published values before this article was written and both match.

The conduit family matters more than people expect, because the wall thickness eats the space. Nine 12 AWG conductors give a fill of 21.80 % in rigid metal at three quarters of an inch, 22.46 % in EMT, 23.56 % in PVC Schedule 40 and 29.27 % in PVC Schedule 80 — a third more fill for the same wires in the same nominal size, because Schedule 80's thicker wall leaves a much smaller bore. Intermediate metal is the roomiest of the five at 20.43 %. None of these fails, but on a run near the limit the choice of pipe is the difference between passing and going up a size.

53, 31 and 40 are not a series — and two conductors get the tightest rule

Read the three limits in order and something looks wrong: 53 %, then 31 %, then back up to 40 %. It is not a mistake and it is not a smooth curve, because the three cases are geometrically different problems. A single conductor in a round pipe is a circle in a circle and can be quite large before it jams — hence 53 %. Two conductors side by side are the worst case for jamming: they can wedge across the bore at a bend and lock, so the code drops the allowance to 31 %. Three or more can shuffle into a triangle or a bundle that rolls around a bend, and the limit relaxes to 40 %.

The calculator picks the limit from the conductor count you type, and the switch is abrupt in a way worth knowing about. Type 2 and the limit shown is 31 %; type 3 and it becomes 40 % while the fill you actually have goes up. The cliff also catches fractional entries: the count is rounded before the limit is chosen, so 2.4 conductors is treated as two and given the 31 % limit, while 2.6 is treated as three and given 40 %. And a count of zero passes at 53 %, which is arithmetically consistent and practically meaningless — an empty pipe is not a design.

The nipple case is the fourth limit and the one most often abused. A short length of conduit no more than 24 in long between two boxes may be filled to 60 %, because there is no meaningful pull and, in a length that short, no meaningful temperature rise. Select it in the tool and the limit jumps to 60 % whatever the conductor count. It is a genuine allowance and not a loophole: 25 in is not a nipple, and a nipple with a bend in it is not a nipple either.

Derating: use the space, lose the current

This is the part the calculator does not do, and it is the part that decides whether the run works. The same heat argument that produced the fill limit produces a second, sharper rule: once more than three current-carrying conductors share a raceway, the code reduces the ampacity of every one of them by a factor that gets worse as the count rises. The steps are in NEC 310.15(C)(1). Four to six conductors keep 80 % of their rating, seven to nine 70 %, ten to twenty 50 %, and it continues down from there. A 12 AWG THHN conductor rated 30 A at 90 °C therefore carries 24 A in a group of six, 21 A in a group of nine, and 15 A in a group of sixteen. The pipe still passes the fill check at sixteen. The circuits in it have halved.

The two counts are not the same count, and mixing them is a common and consequential error. For the fill limit, every conductor in the pipe counts, equipment grounding conductors included — they occupy space like anything else. For the ampacity adjustment, only current-carrying conductors count: a grounding conductor never counts, and the neutral of a balanced multi-wire circuit generally does not either, though a neutral that carries harmonic current does. So a pipe with sixteen conductors in it might contain twelve that are current-carrying, and the adjustment factor depends on the twelve while the fill depends on the sixteen. Getting that backwards either derates a circuit that did not need it or, far worse, fails to derate one that did.

There is a further constraint that keeps the arithmetic honest. The 90 °C rating a conductor derates from is not the current it is allowed to deliver: terminations on ordinary devices are listed at 60 °C or 75 °C, and the circuit is limited by the lower of the derated ampacity and the termination rating. Separately, small copper conductors are capped by their own rule on overcurrent protection regardless of any of this. The practical consequence is that the headroom the 90 °C column appears to offer is mostly there so that you have something to derate from, not so that you can use it.

Outside the United States, none of these numbers apply

The NEC is a United States document, published by the NFPA and adopted state by state and city by city with local amendments. Everywhere else, the family that matters is IEC 60364, implemented nationally: NF C 15-100 in France, the DIN VDE 0100 series in Germany, the REBT in Spain, the RTIEBT in Portugal, CEI 64-8 in Italy, BS 7671 in the UK. None of them contains the 53 / 31 / 40 numbers, none of them tabulates conduit fill in square inches, and the conductor sizes are cross-sections in square millimetres rather than gauge numbers. Conduit itself is a different product line: sized by outside diameter to the EN 61386 series — 16, 20, 25, 32, 40, 50, 63 mm — where the American series is a trade size in inches that is not the bore or the outside diameter of anything in particular.

What does transfer is the arithmetic, because a fill fraction has no units. Take the conduit's internal diameter, square it and multiply by π/4 to get the bore area. Take each cable's overall diameter from its data sheet — the overall diameter, not the conductor cross-section, because the sheath is what occupies the pipe — square it, multiply by π/4, and sum. Divide the second by the first. A 20 mm conduit with a 1.5 mm wall has a 17 mm bore and 226.98 mm² of area; nine cables of 3.5 mm overall diameter occupy 86.59 mm², which is 38.15 % fill. Move the same nine cables into a 25 mm conduit and the figure drops to 22.78 %. The number is meaningful; the limit you compare it against is not the NEC's, and has to come from the rule that applies where you are working.

The grouping penalty exists on both sides of the Atlantic, which is the reassuring part. IEC 60364-5-52 carries reduction factors for circuits bunched together in a conduit or trunking, tabulated by the number of circuits rather than the number of conductors, and they fall off in the same shape as the NEC's: the more circuits share the enclosure, the less each may carry. The two systems disagree about the numbers, about what is counted, and about the reference installation method a rating starts from. They agree completely on the physics. If you take one idea away from this article, take that one — the pipe is a thermal problem before it is a spatial one, in every country.

What the calculator does not check

It takes one conductor size and one conductor count, which is not what a real pull looks like. A branch circuit is typically two or three conductors of one size plus a smaller equipment grounding conductor, and a feeder often has a reduced neutral. There is no way to enter that mixture, so the honest workaround is to run the tool for each size, add the conductor areas by hand, and compare the total against the same limit — the limit itself depends only on how many conductors there are in total, not on their sizes.

It only knows THHN and THWN. Those are the common building wires in the United States and the reason the tool can get away with one table, but XHHW, RHW, USE and everything else occupy different areas for the same copper, sometimes noticeably more. Nor does it apply the jam ratio — the rule that three conductors of the same size can wedge in a bend when the conduit's inside diameter divided by the conductor's outside diameter lands in a narrow band around three. And it says nothing about pull tension, bend count, box fill, or whether the conductor is the right size for the load in the first place, which is a different calculation and a different tool.

Frequently asked questions

Does the equipment grounding conductor count towards the fill?

Yes for fill, no for the ampacity adjustment, and the distinction is the single most useful thing in this article. Fill is a question of occupied area, so anything in the pipe occupies area — grounding conductors included, whether bare or insulated, and whether or not they carry current in normal operation. The ampacity adjustment is a question of heat generated, so only conductors that actually carry load current count; a grounding conductor carries current only during a fault, which lasts milliseconds. The same reasoning exempts the neutral of a balanced multi-wire circuit, where the neutral only carries the imbalance — with the well-known exception of a neutral serving loads that produce harmonic currents, which does carry current and does count.

Why is the limit for two conductors lower than for three?

Because two round objects in a round tube are the worst arrangement for jamming, and jamming is a pulling problem rather than a heat one. Two conductors of similar size can turn broadside-on at a bend and wedge across the bore, and once wedged they do not slide — the harder you pull the tighter they lock, and the sheath tears. Three or more cannot form that particular wedge as easily: they fall into a triangle or a loose bundle that rolls round the bend. A single conductor has nothing to jam against and gets the most generous allowance of all at 53 %. So the sequence 53, 31, 40 is not a curve at all, it is three separate answers to three separate geometric questions, which is why the tool switches limits discontinuously as you change the count.

Can I use this tool for a European installation?

To understand the principle, yes. To size a run, no. Everything in it is American: the conduit series is trade sizes in inches, the conductor areas are THHN and THWN, and the limits are the NEC's. A European installation is governed by the national implementation of IEC 60364, uses conduit sized by outside diameter in millimetres to the EN 61386 series, and specifies cable by cross-section in square millimetres. What does carry over is the method, because a fill fraction is a ratio and has no units: compute the bore area of your conduit, sum the areas of the cables from their overall diameters, and divide. That figure is directly comparable to a fill limit — just not to the NEC's fill limit. Take the number to the rule that applies where the cable is going, and remember that the equivalent of the derating question exists there too, as grouping reduction factors in IEC 60364-5-52.

How do I handle a pull with more than one conductor size?

By hand, because the tool takes one size at a time. Run it once per size to read off the total conductor area for that group, add the groups together, and compare the sum against the conduit's internal area. The limit does not change with the mixture: it is chosen purely by how many conductors there are in total, so three of one size plus one smaller grounding conductor is four conductors and gets the 40 % limit, not some blend. The trap in the arithmetic is forgetting the smallest conductor because it looks negligible — a grounding conductor in a tight run is often exactly the thing that pushes a marginal fill over the line, and it is the one people leave out of the count.

If a run passes the fill check, is the conduit correctly sized?

No. Fill is one gate among several, and it is the easiest one to pass. A run also has to survive the pull, which depends on its length, the number and radius of its bends and the lubricant used, and none of that is in a fill percentage. It has to satisfy the ampacity adjustment for however many current-carrying conductors it holds, which may force a larger conductor and therefore a larger conduit than the fill check suggested. It has to avoid the jam ratio if it carries exactly three same-size conductors. The boxes at each end have their own fill rules. And the conductor itself has to be sized for its load and its voltage drop before any of this starts, which is where a run should be designed from rather than backwards from the pipe. Passing the fill check means the conduit is not disqualified; it does not mean it is right.