Cable Pulling Physics — Why Every Bend Multiplies Tension (Capstan Equation)

Cable installation destroys more cables than overload, short circuits, and rodents combined. Torn jackets, crushed insulation shields, stretched conductors — all caused by pulling forces that exceeded safe mechanical limits. The frustrating part: every one of these failures was preventable with a calculation that takes five minutes.

Last year a contractor on a data centre project in Southeast Asia pulled 3 × 185mm² XLPE cables through a 100-metre underground duct bank with four 90° sweeps. The conduit fill was fine at 28%. The cable size was correct for the load. But the pulling force required at the fourth bend exceeded the cable manufacturer's maximum tension by 340%. The outer sheath ruptured 60 metres into the pull. Three drums of cable — scrapped.

The problem wasn't the cable or the conduit. It was the bends. Each bend multiplied the tension exponentially, and nobody ran the numbers.

Why Cable Pulling Isn't Just "Pull Harder"

Cable pulling creates four distinct mechanical stresses, and each one can independently damage the cable:

1. Longitudinal tension — the axial force pulling the cable through the conduit
2. Sidewall bearing pressure (SWBP) — radial force where the cable presses against conduit walls at bends
3. Bending strain — deformation as the cable follows curved sections
4. Friction — resistance between the cable jacket and the conduit interior along straight runs

Get any single stress above the safe limit and the cable is damaged, even if the other three are fine. Most engineers check conduit fill and stop there. Fill percentage tells you whether the cables fit. It tells you nothing about whether you can get them in without damage.

The Capstan Equation: Where Bends Become Multipliers

The critical insight in cable pulling physics is that bends don't add to tension — they multiply it. This relationship is described by the capstan equation (also called Euler-Eytelwein formula), the same physics that lets a sailor hold a ship with a rope around a bollard.

Where:

• T_out = tension on the pulling side of the bend (N)
• T_in = tension on the entry side of the bend (N)
• μ = coefficient of friction between cable jacket and conduit wall
• θ = bend angle in radians (90° = π/2 = 1.571 rad)
• e = Euler's number (2.71828...)

IEEE 1185, Section 4.2 — Pulling tension calculations at bends

The exponential makes this equation brutal. With a typical friction coefficient of μ = 0.5:

A route with three 90° bends doesn't triple the tension. It multiplies by 2.19 × 2.19 × 2.19 = 10.5× the entry tension. Four 90° bends: 23×.

Straight Sections: Weight, Friction, and Slope

Between bends, straight sections add tension linearly based on cable weight, friction, and inclination:

Where:

• w = cable weight per unit length (N/m) — multiply kg/m by 9.81, then by number of cables
• L = section length (m)
• α = inclination angle from horizontal (positive = uphill pull)

The first friction term (w × L × μ × cosα) is always present. The gravity term (w × L × sinα) adds tension for uphill pulls and subtracts for downhill pulls.

Sidewall Bearing Pressure: The Invisible Killer

Pulling tension tells you whether the cable can handle the axial load. But at every bend, the cable presses against the conduit wall with a radial force called sidewall bearing pressure (SWBP). Excessive SWBP crushes the insulation shield and deforms the conductor cross-section.

Where:

• T = pulling tension at the bend (N)
• r = bend radius (m)

AEIC CS8, Section 7 — Maximum allowable sidewall bearing pressure

Typical SWBP limits:

SWBP often governs before tension does, especially on tight-radius bends. A cable that passes the maximum tension check can still fail the SWBP check at a tight 90° elbow.

The Jam Ratio: When 3 Cables Won't Fit (Even Though They Should)

Conduit fill says the cables fit. Physics says they jam. This happens specifically with three cables of the same diameter in a circular conduit when the ratio of conduit internal diameter (D) to cable external diameter (d) falls in a critical range.

NEC, Chapter 9, Table 1, Note 6 — Jam ratio awareness

At D/d ≈ 3.0, three cables form an equilateral triangle that exactly spans the conduit diameter. They wedge against each other and the conduit wall. No amount of tension, lubrication, or mechanical advantage will free them — the geometry is physically locked.

This is one of the most frustrating problems in cable installation because:

• The conduit fill calculation says it's fine (typically around 35-40%)
• The conduit meets code requirements
• Installation proceeds normally until the cables suddenly lock
• Backing out and pulling again doesn't help — the geometry repeats

Solutions when D/d is in the jam zone:

1. Upsize conduit — push D/d above 3.2 (cables can roll past each other)
2. Downsize conduit — push D/d below 2.8 (cables stack in a clover pattern)
3. Use 2 or 4 cables instead — the jam geometry only applies to exactly 3
4. Select cables with different OD — even a slight diameter difference breaks the symmetry

Pull Direction Matters

Here's something that surprises most engineers: the direction you pull from can halve the required tension. The capstan equation multiplies tension at each bend. Since the multiplication compounds sequentially, the order in which the cable encounters bends determines the final tension.

Rule of thumb: Pull towards the end that has fewer remaining bends. Each subsequent bend multiplies a larger number, so you want bends early in the route (where tension is still low) rather than late (where it's already high).

Consider a route: 50m straight → 90° bend → 30m straight → 90° bend → 20m straight.

Pulling left to right (bends in the middle and near the end): the first long section builds up tension, and both bends multiply that accumulated tension.

Pulling right to left (bends near the start): the bends multiply a smaller initial tension, and the long straight adds only a linear component afterward.

The difference can be 30-50% in final tension. For borderline installations, simply reversing the pull direction can mean the difference between a successful pull and a damaged cable.

Pulling Lubricant: The Cheapest Insurance

Cable pulling lubricant (also called pulling compound) reduces the friction coefficient μ from 0.5–0.8 down to 0.2–0.3. Since μ appears in the exponent of the capstan equation, this reduction has an enormous effect.

For a route with 270° total bends:

• Without lubricant (μ = 0.5): multiplier = e^(0.5 × 4.712) = 10.6×
• With lubricant (μ = 0.25): multiplier = e^(0.25 × 4.712) = 3.2×

Lubricant reduced the tension by a factor of 3.3. A $50 bucket of pulling compound saved three drums of cable on the data centre project I mentioned at the start of this article.

Maximum Tension Limits

The pulling tension must not exceed the weakest link in the cable construction. For copper and aluminium conductors:

Where:

• k = conductor stress limit: 70 MPa for copper, 52 MPa for aluminium
• A = cross-sectional area of one conductor (mm²)
• n = number of conductors being pulled simultaneously

IEEE 1185, Table 3 — Maximum conductor tension limits

For a 3 × 185mm² copper pull: T_max = 70 × 185 × 3 = 38,850 N (38.85 kN).

If your calculated pulling tension exceeds T_max, you have three options:

1. Reduce bend angles or use larger bend radii
2. Apply pulling lubricant
3. Split the pull into sections with intermediate pulling points

A Complete Worked Example

Scenario: Pull 3 × 95mm² single-core XLPE copper cables through 80mm PVC conduit.

Route: 30m horizontal → 90° bend (r = 0.6m) → 15m uphill at 30° → 90° bend (r = 0.6m) → 20m horizontal

Cable data: OD = 16.2mm, weight = 1.16 kg/m per cable (3.48 kg/m total = 34.1 N/m)

Conduit data: ID = 80.5mm, μ = 0.5 (unlubricated PVC)

Step 1: Check jam ratio D/d = 80.5 / 16.2 = 4.97 — well above 3.2, no jam risk.

Step 2: Check conduit fill Cable area = 3 × π × (16.2/2)² = 618.8 mm² Conduit area = π × (80.5/2)² = 5,089.6 mm² Fill = 618.8 / 5,089.6 = 12.2% — well within 40% limit.

Step 3: Calculate tension (segment by segment)

Starting tension = 0 N (cable weight feeds from drum):

Segment 1 — 30m horizontal:

Bend 1 — 90°:

Segment 2 — 15m uphill at 30°:

Bend 2 — 90°:

Segment 3 — 20m horizontal:

Step 4: Check against limits

Maximum tension: T_max = 70 × 95 × 3 = 19,950 N Calculated tension: 3,842 N (19.3% of limit) — PASS

SWBP at worst bend (Bend 2): 3,501 / 0.6 = 5,835 N/m Limit for XLPE: 4,400 N/m — FAIL

Despite the tension being well within limits, the SWBP at Bend 2 exceeds the safe limit by 33%. Solutions:

• Use a long-radius bend (r = 0.9m): SWBP = 3,501 / 0.9 = 3,890 N/m — PASS
• Apply lubricant (μ = 0.3): recalculate — tension at Bend 2 drops to ~1,100 N, SWBP = 1,833 N/m — PASS

This is exactly why you need the calculation. Fill was fine. Tension was fine. SWBP would have damaged the cable.

Key Takeaways

1. Bends multiply tension exponentially — each 90° bend roughly doubles the pulling tension at μ = 0.5
2. Check all four limits — conduit fill, pulling tension, SWBP, and jam ratio. Passing three out of four still fails
3. Pull direction matters — the optimal direction can reduce maximum tension by 30-50%
4. Lubricant is essential on any route with more than 180° total bends
5. SWBP often governs before tension does, especially with tight-radius bends
6. The jam ratio catches exactly 3 cables when D/d is between 2.8 and 3.2

Related Resources

• Conduit Fill: Everyone Gets It Wrong — Conduit fill limits that interact with cable pulling feasibility
• Cable Sizing Guide: The Complete Method — Size the cable before planning the pull
• Aluminium vs Copper: The Comeback Story — How conductor material affects pulling tension limits
• Mining Cable Pulling Feasibility Study — Full worked example at a copper-gold mining operation
• View all worked examples →