Torsion of Circular Shafts
A twisted shaft works hardest at its surface and barely at all at its center.
Before this Lesson
- Strain, Hooke's Law, and Poisson's Ratio
- Composite Sections · Statics
- Power and Efficiency · Dynamics
Draw a grid on a thick rubber rod: lines running along it, and rings around it. Hold one end and twist the other. The lines wind into gentle spirals, and every little square on the surface leans over into a diamond. The rings stay rings, the same size and the same distance apart. That's all you can see from outside, and it's enough to tell you where a shaft works hardest, how far it twists, and why a truck's drive shaft is a hollow tube.
Every slice turns a little more than the last
The rings carry the key fact. Each cross-section of a round shaft stays flat and round and turns as a rigid disk, like a coin on a skewer. With one end fixed, each slice turns a little more than the one before, and the free end turns through the whole angle of twist, .
Follow a line along the surface. Its far end, on the rim at radius , swings along an arc . Spread over the length , that tilts the line, and every little square on it, by a small angle : the shear strain of Strain, Hooke's Law, and Poisson's Ratio. For a small angle, , so:
A line deeper inside, at a smaller , swings a shorter arc and tilts less, and at the center it doesn't tilt at all. So the strain grows in proportion to the distance from the center. The Handbook's Torsional Strain entry says so, writing , with the twist per unit length in place of .
Hooke's law in shear, , ties the stress at each point to the strain there, and is the same everywhere in the shaft. The stress acts along the circle through each point, the way the slices slide.
This needs a round shaft. A twisted square bar's cross-sections warp out of their plane, and none of it holds.
Adding up the torque
Stress follows strain, so it's zero at the center and grows in a straight line to its peak at the surface. Those stresses on a cut have to add up to the torque the shaft carries. Take a small patch of area at a distance from the center. The patch carries a force on an arm , a torque of .
With :
There's , showing up twice. Material farther out is strained more, so it pushes harder, and it pushes on a longer arm. That makes the sum , the polar moment of inertia . A wider shaft also has more area out at those large radii, which is why grows with the fourth power of the radius.
From Composite Sections: for a solid shaft and for a hollow one. The Handbook's Statics area table prints them with radii, as , and a diameter in that form lands 16 times too high.
Solve that line for and put it back into . The cancels, leaving the formula in the Handbook's Torsion entry:
It gives the stress at any distance from the center, and the largest at the surface, where , not . The Handbook states it for solid shafts and thick-walled tubes, which covers every shaft here.
is the internal torque at the section. Finding it is the statics half, and in a shaft driven at one end and loaded at the other, it's the same all along. Keep one unit family: kip-in., in., and in⁴ give ksi, and N·mm, mm, and mm⁴ give MPa.
Hollow shafts put the steel where the stress is
Look at the stress picture again. The steel near the center is barely stressed, and its force has almost no arm, so it adds weight and little strength. Spend that steel on a wider ring instead, and the shaft gets much stronger. That's the drive shaft's secret.
A tube's is the outer circle's minus the hole's, : fourth powers first, then subtract. A wall thickness comes off both sides, so .
The stress still grows with , but the line starts at the inner wall instead of at zero. gives both surfaces, with outside and inside, so .
Those two sections hold the same steel. The tube's is 3.5 times the solid's, so its peak stress is 2.73 ksi against 6.37 ksi, and at the same stress it could carry 2.33 times the torque.
Torque from power
A turning shaft carries power , as in Power and Efficiency. The Handbook prints it under Rotating Machines. A motor's nameplate gives power and speed, so turn it around, , and take care with the units, from the Handbook's Units and Conversion Factors table:
- Horsepower times 550 gives ft-lb/s. Kilowatts times 1,000 give watts, or N·m/s.
- Rpm times gives rad/s.
- The torque comes out in ft-lb or N·m. Times 12 gives in.-lb, and times 1,000 gives N·mm, ready for .
The thinnest shaft that's strong enough
For a solid shaft, put and into , and the stress collapses into one handy form. The Handbook doesn't print it, but it's one substitution away:
A thinner shaft is more stressed, so this diameter is a floor: if a question offers sizes, take the first one above it. If it gives power instead of torque, find the torque first.
How far it twists
The torque sum already handed over the twist. Solve for , and you have the Handbook's angle of twist, under Torsional Strain:
It holds where the torque, the material, and the section stay the same along the length . Set it beside axial loading, and the two match term for term:
| Axial | Torsion | |
|---|---|---|
| Internal load | ||
| Stress | ||
| Modulus | ||
| Deformation | ||
| Stiffness |
Two things to watch. Twisting is sliding, so the modulus is , the modulus of rigidity in Table 1 - Typical Material Properties: 11,500 ksi or 80,000 MPa for steel, and 3,800 ksi or 26,000 MPa for aluminum. And comes out in radians, as the Handbook says, so multiply by for degrees.
A shaft is a torsional spring
Rearrange the way Axial Deformation rearranged :
is the torque per radian of twist. The Handbook calls the torsional stiffness, and its Torsional Vibration entry prints . Given a twist, the stiffness gives the torque. Given a twist limit, it gives the least the shaft can have, and so its diameter:
Put the limit in radians before it goes in. A shaft with both a stress limit and a twist limit needs the larger of the two diameters.
Segments in a row
A stepped shaft is torsional springs in a row, like a stepped bar. Each segment twists by its own share, and the shares add:
Each is that segment's internal torque, found by cutting, as in Axial Stress: keep the piece away from the wall, and the torque at the cut balances the torques on that piece, each with its sign.
The case to know cold is a shaft fixed at one end, with the torque part way along. Cut beyond the torque, and the piece you keep has no torque on it, so that segment carries nothing and doesn't twist. It still turns, riding along with the point where the torque acts, so the free end turns exactly as far as that point.
Changing one thing at a time
For a solid shaft, put into both formulas, and the powers of the diameter show:
The stress goes with , because brings four powers of the diameter and gives one back. The twist goes with , from alone. Length is in the twist but not the stress: a longer shaft twists more, but no slice of it works any harder. To compare two shafts, write new over old and cancel whatever didn't change.