How the two differ in kind
Two separate limits govern what can go into a conduit run, and each will stop you for its own reason.
FILL is a percentage of the raceway's internal cross-sectional area occupied by the conductors, with the permitted figure falling as the number of conductors rises. It exists for two reasons: heat dissipation, since bundled conductors in a full raceway cannot shed heat, and physical practicality, since conductors packed beyond a certain proportion cannot be pulled without damage. It is a property of the conduit and the conductors alone, so a run of two metres and a run of two hundred have identical fill.
PULLING TENSION is the force needed to drag the conductors through, and it is entirely a property of the ROUTE. It grows with length, as friction accumulates along the run, and it grows far more sharply with BENDS — because of the capstan effect, the same principle that lets a rope round a bollard hold a ship. A conductor going round a bend presses against the inner wall of that bend with a force proportional to the tension already in it, and friction acts on that pressure — so the tension coming out of a bend is the tension going in MULTIPLIED by a factor, not increased by an amount. Two ninety-degree bends multiply twice.
The consequences follow directly. A conduit comfortably within its fill limit can be unpullable because of its bends. Conversely, a short straight run at the fill limit pulls easily. And the two limits are not interchangeable: a bigger conduit reduces both, which is why it is the usual answer, but the fill calculation alone will not tell you that a run needs a pull box.
The failure mode matters too. A pull that exceeds the safe tension does not usually stop; it continues, and the damage is to the conductor — insulation scraped against the raceway or stretched beyond its limit, and a fault that appears later rather than at the time.
The factors that actually differ
Show
| Fill percentage | Pulling tension | |
|---|---|---|
| What it limits | How much conductor area may occupy the raceway. | How much force the pull will require, and therefore whether it can be done safely. |
| Depends on length | No. A two-metre run and a two-hundred-metre run have identical fill. | Yes, directly — friction accumulates along the run. |
| Depends on bends | No. | Enormously. Tension is MULTIPLIED at each bend, not increased. |
| Why the limit exists | Heat dissipation, and so that a pull is physically possible. | To avoid damaging the conductor — insulation abrasion and stretching. |
| What happens when exceeded | The pull becomes very difficult or impossible; the conductors also derate. | The pull often still succeeds, and the conductor is damaged. The fault appears later. |
| The usual fix | A larger conduit, or fewer conductors per raceway. | Pull boxes to break the run, fewer or gentler bends, lubricant, and a larger conduit. |
| Code involvement | An explicit percentage, in tables. | A limit on total bend angle between pull points, plus the cable manufacturer's maximum tension and sidewall pressure. |
| Direction of pull | Irrelevant. | Matters. Pulling from the end with the bends closest to it is easier than pulling them at the far end. |
| Lubricant | No effect on the number. | Substantial — it reduces the friction coefficient, which reduces both the length and bend terms. |
| Which binds first | Short runs and congested raceways. | Long runs and anything with several bends — and it binds without warning from the fill figure. |
Which one, and when
Choose fill percentage when…
- Selecting a conduit size for a given set of conductors — this is where every design starts.
- Checking compliance, since fill is an explicit code requirement.
- Assessing whether an existing conduit has room for an additional circuit.
- Any run where heat is a concern, since fill and derating are related.
Choose pulling tension when…
- Long runs, where accumulated friction is substantial.
- Any run with multiple bends, where the multiplication effect dominates.
- Large conductors, which are heavy and stiff and pull hard.
- Planning pull box positions, which is what the calculation is actually for.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Why do bends matter so much more than length?
- Because of the capstan effect: the tension leaving a bend is the tension entering it MULTIPLIED by a factor that depends on the friction coefficient and the bend angle, rather than increased by a fixed amount. The mechanism is that a conductor under tension going round a bend presses against the bend's inner wall, and the friction force resisting the pull is proportional to that pressure — which is itself proportional to the tension already in the conductor. So tension compounds through successive bends in the same way that interest compounds. A run with several ninety-degree bends can require many times the tension of a straight run of the same length, which is why codes limit the total bend angle between pull points rather than limiting length, and why the position of the bends along the run also matters.
- Why does fill fall as the number of conductors rises?
- Because both reasons for the limit get worse. On heat: more current-carrying conductors bundled together means each one has less access to still air and they warm each other, so the ampacity of each is derated — and a raceway packed with conductors is the extreme case of that. On practicality: several conductors being pulled together do not pack efficiently, they cross over one another within the raceway, and the pull becomes far harder than the simple area ratio suggests. A single conductor in a conduit can occupy a much greater proportion of the area than three or more can, which is exactly what the tables reflect. Fill is also calculated on the conductors' actual outside dimensions rather than their nominal sizes, since insulation thickness varies by type.
- What actually gets damaged in an over-tension pull?
- The insulation, usually, and the damage is not visible from either end of the run. Excessive tension stretches the conductor and can reduce the insulation's thickness where it is drawn over a surface; excessive SIDEWALL PRESSURE — the force per unit length pressing the conductor against the inner wall of a bend — crushes or abrades the insulation at that bend specifically. Both leave a conductor that tests fine on installation and fails later, sometimes years later, at a point nobody can reach. That is why cable manufacturers publish a maximum pulling tension and a maximum sidewall pressure, why tension monitoring is used on significant pulls, and why the honest response to a pull that is going hard is to stop and add a pull point rather than to find a bigger winch.
- What is sidewall pressure and why is it a separate limit?
- The force per unit length that a conductor exerts against the inside of a bend as it is pulled round it — equal to the tension at that point divided by the bend's radius. It is a separate limit from total tension because it depends on the bend RADIUS as well as the tension: the same tension round a tight bend produces far higher local pressure than round a sweeping one, and it is the local pressure that crushes insulation. That is why a large-radius bend is so much kinder than a standard elbow, why sweeps are specified on large-conductor runs, and why a pull can be within its tension limit and still damage the cable at one particular bend. Both limits are published by the cable manufacturer and both are checked.
- How are pull boxes decided?
- By the total bend angle permitted between them and by the tension calculation, and the first is the code requirement while the second is the engineering one. Codes limit the cumulative bend angle in a run between pull points — commonly a total of a few right angles — which is a straightforward rule and is the minimum. Beyond it, the tension calculation identifies where the force would exceed the cable's rating and therefore where a box is needed regardless of the angle count, which on a long run with large conductors can be sooner than the angle limit implies. Box positions also want to be accessible, since a pull box in a ceiling void above a fixed installation is a box nobody can use for the next pull, which defeats the purpose of having it.
- How much does lubricant help?
- Substantially, because it reduces the friction coefficient, and that coefficient appears in both the length term and, exponentially, in the bend term. Halving the friction roughly halves the straight-run contribution and reduces the bend multiplication considerably more, so lubricant can be the difference between a pull that is safe and one that is not. The conditions are that it must be compatible with the cable's jacket and with the raceway — some lubricants degrade certain jacket materials over time — and that it has to be applied continuously as the cable enters rather than poured in at the start, since lubricant that stays at the entry does nothing at the far bends. It is not a substitute for a pull box on a run that needs one; it is what makes a marginal pull comfortable.
- Does it matter which end you pull from?
- Yes, and choosing well can be free. Tension multiplies through each bend, so the bends closest to the PULLING end are multiplying a tension that has already accumulated through everything before them — while bends near the feeding end multiply a small tension. It follows that pulling from the end nearest the bends is easier than pulling them at the far end of a long run: the same route can require significantly different tensions depending on the direction. Where a run has its bends concentrated at one end, pulling from that end is the correct choice. It is also why the calculation is performed for the intended direction rather than generically, and why a run that failed in one direction is sometimes worth attempting in the other before adding a box.
- Does any of this apply to cable tray?
- Neither limit applies in the same way, which is one of tray's practical advantages. Cables are LAID into a tray rather than pulled through it, so pulling tension is largely irrelevant — there is no raceway wall to drag against and no bend to multiply tension round. And tray fill is governed by different rules based on the sum of cable diameters and on layer depth rather than by a percentage of cross-sectional area, because the concern is heat and the tray's load capacity rather than the possibility of a pull. That is a large part of why tray is chosen for routes carrying many circuits: adding a circuit is a matter of laying a cable in rather than calculating whether it can be pulled, and the route's bends stop being a constraint on what can be installed.