The arithmetic is the same, the limit is not
Fill is a cross-section problem and nothing more complicated than that. Each cable occupies pi times its diameter squared over four. The pathway offers pi times its inside diameter squared over four. Divide one by the other and you have a percentage. That much is identical whether the cables carry mains power or data.
What differs is the limit you compare the percentage against, and the reasoning behind it. For power conductors in a raceway the conventional limits come out of the electrical code and are tied to conductor counts. For communications pathways the figure comes from cabling standards and manufacturer guidance and is not automatically the same number. This page therefore takes the limit as an input rather than baking one in, and you should fill that field from the specification you are actually installing to. If the question you have is about power conductors in conduit, that is a different table and a different page: the conduit fill calculator handles conductor sizes and raceway types directly.
Why the limit is well below full
A percentage in the region of forty sounds absurdly conservative until you have pulled a bundle through a bend. The limit is not about whether the cables occupy the space. It is about four things that only appear during and after the pull.
| What the headroom buys | What happens without it |
|---|---|
| Room for the bundle to twist and lay up | Cables bind against each other and tension climbs steeply |
| Room at the inside of every bend | The bundle compresses, jackets deform, pairs shift |
| Tolerance for the pathway not being perfectly round or straight | A crushed section becomes a hard stop partway through |
| Capacity for later additions | The next cable means opening a wall |
Data cable punishes deformation in a way power conductors do not. A twisted pair transmission line depends on the geometry of the twist and the spacing between conductors. Squash it, kink it, or pull it hard enough to stretch it and the electrical performance changes permanently, and the symptom is a link that negotiates a lower speed or throws errors under load rather than one that visibly fails.
Bend radius, and why the pull figure is the one that matters
Every data cable has a minimum bend radius, expressed as a multiple of its own outside diameter, and there are usually two of them: a tighter figure for the installed cable at rest and a larger one for the cable under pulling tension. The pulling figure is the constraint on the pathway, because that is the condition the cable is in when it goes around the sweep.
This is where tight sweeps and the sharp fittings that make conduit runs convenient become a problem. A ninety degree turn made with a short radius fitting can be inside the cable minimum on its own, which means the cable is being damaged at the moment it is installed by somebody doing everything else right. Long sweeps, or a pull box at the corner, are the answer, and pull boxes have the additional virtue of splitting a long run into two shorter pulls.
The jam, which fill percentage does not predict
There is a specific and infuriating failure that the fill arithmetic says nothing about. Pull three cables of the same diameter through a conduit whose inside diameter is about three times theirs, and at a bend they can arrive side by side across the bore and wedge. The bundle stops. Pulling harder makes it worse, because it drives the wedge tighter. The fill percentage in that situation is comfortable, which is exactly why the failure surprises people.
The ratio to watch is the pathway inside diameter divided by the cable outside diameter, and the awkward region is roughly 2.8 to 3.2. The calculator flags it when you have three identical cables, since that is when it applies. Moving out of the band in either direction fixes it: a larger pathway, or a different cable count, or accepting that the three cables go in as two pulls.
Leaving the pathway usable
A conduit that is full is a conduit that has done its job once. The habit that separates installations people are glad of from ones they curse is treating spare capacity and a pull string as part of the specification rather than as good practice. Size the pathway for what goes in it plus what you would be annoyed not to be able to add, pull a string alongside the first bundle, and tie the string off at both ends so the first pull does not carry it away.
Label both ends while the ends are visible. A pathway whose far end nobody can find is functionally a pathway that does not exist, and the person who needs to find it will not be you. This is the cheapest item on any structured wiring job and the one most often skipped, and it is the reason the ethernet drop calculator counts service loops as material rather than as slack: both are money spent now to keep a future option open. Where the pathways converge, the structured wiring panel planner works out whether the enclosure at the end of them is big enough.
Questions people ask
How many network cables fit in a given conduit?
It depends on the outside diameter of the cable you are actually using, which varies far more than people expect. A plain unshielded cable, a shielded one and a higher category one of the same nominal type can differ by a third in diameter, and area goes with the square of diameter, so a third fatter is nearly eighty percent more area each. Take the outside diameter from the datasheet for the reel you bought rather than from a general figure, and enter the inside diameter of the pathway rather than the trade size, because trade sizes are names and not measurements.
What fill percentage should I use for data cable?
Take it from the standard or manufacturer guidance you are installing to rather than from a figure remembered from power work. The conventions for communications pathways come from cabling standards and are not automatically the same numbers used for conductors in a raceway under the electrical code, and both change between editions and between jurisdictions. That is why this page takes the percentage as an input. What is safe to say generally is that the limit is well under a hundred for reasons that have nothing to do with whether the cables physically fit, and that a pathway filled to its stated maximum has no room for the cable you will want next year.
Does conduit have to be used at all for low voltage cable?
That is a question for the code adopted where you are and for the authority having jurisdiction, and the answer varies. What is generally true is that cable installed inside a wall, above a ceiling or in an air handling space has rating requirements, that those requirements are real and enforced, and that they are separate from the question of whether a pathway is required. Beyond compliance, a pathway is worth installing on its own merits at the places where you would otherwise have to open finished work: between floors, into a loft, out to an outbuilding, and anywhere the route crosses something that will be sealed.
Why did my pull stop halfway with plenty of room left?
The two usual causes are a jam and a bend. A jam happens when several cables of the same diameter wedge across the bore, most notoriously three cables in a pathway about three times their diameter, and the fill percentage looks fine when it happens. A bend problem happens when the accumulated tension around several turns exceeds what you can apply from one end, which is why every bend in the run makes the pull disproportionately harder. Pulling harder is the wrong response to both. Back off, add lubricant intended for cable, pull from the other direction, or split the run at a pull box.
Can data cable share a conduit with mains wiring?
That is governed by code and by the authority having jurisdiction, it differs by jurisdiction and by edition, and anything sharing a pathway or a box with line voltage is licensed work. This page gives no procedure for it and no separation distance. What is worth knowing independently of any code question is that running data alongside power invites induced noise, that the cable ratings for the two are different, and that on any job where the question comes up the correct move is to ask the electrician who is already there rather than a web page.