Torsion testing machines: what they measure, the standards and how to choose Nm capacity

Definition: A torsion testing machine twists a specimen between a motor-driven head and a fixed head carrying a torque transducer, recording torque (N·m) against angle of twist (degrees) to give the shear modulus G, torsional yield or proof torque, maximum torque, angle at fracture and, for wire, the number of twists withstood before failure.

1. What a torsion test measures

A tensile test pulls; a torsion test twists. The record is a torque–angle curve, the counterpart of the force–extension curve on a UTM, and from it the software extracts:

• Shear modulus G (modulus of rigidity) — slope of the linear part of the curve, converted to shear stress against shear strain. Steel lies around 79–81 GPa, aluminium alloys about 26 GPa, copper about 45 GPa.
• Torsional yield or proof torque — torque at a defined permanent twist, the equivalent of R_p0.2.
• Maximum torque — the peak of the curve; a ductile bar keeps twisting well beyond it because the section does not neck.
• Angle of twist to fracture — ductility, in degrees or full turns over the gauge length.
• Number of twists (wire) — the IS 1717 simple torsion test counts complete 360° turns to fracture, or to the count the product specification demands, and records the fracture appearance.

Shafts, fasteners, springs and drill bits carry torsion in service. For drawn wire and welded joints the test also exposes seams, laps and lack of fusion that a tensile test can pass.

2. The formulas and a worked example

For a solid round bar of diameter d under torque T, the shear stress at the surface is

τ = 16 T / (π d³)

The elastic angle of twist over a gauge length L is

θ (rad) = T L / (G J)   with   J = π d⁴ / 32 (solid round)   or   J = π (D⁴ − d⁴) / 32 (tube, outside D, bore d)

so the shear modulus from a measured elastic slope is

G = (ΔT / Δθ) × L / J

Keep units consistent: T in N·mm, d and L in mm, G in MPa, θ in radians (× 57.3 for degrees). 1 N·m = 1000 N·mm and 1 kgf·m = 9.807 N·m; quote torque in N·m, as the standards and calibration certificates do.

Worked example. A 10 mm alloy steel bar, 100 mm gauge length, twisted to 30 N·m. J = π × 10⁴ / 32 = 981.7 mm⁴. Surface shear stress τ = 16 × 30 000 / (π × 1000) = 153 MPa, comfortably elastic. With G = 80 000 MPa, θ = 30 000 × 100 / (80 000 × 981.7) = 0.0382 rad = 2.19°. The whole elastic part of the test occupies about two degrees; an encoder resolving 0.01° divides that into 200 steps, motor steps through a gearbox with backlash do not.

The formulas are elastic. Once the surface yields, the stress redistributes and the fully plastic torque of a solid round bar is 4/3 of the first-yield torque — a factor used in section 5.

3. What the standards ask for

IS 1717 — Metallic materials, wire, simple torsion test (identical to ISO 7800). A straight wire is gripped with a free length that is a multiple of the diameter d: 100 d for wire from 1 mm up to 5 mm, 200 d for finer wire, 50 d for heavier wire. A small tensile force — enough to keep the wire straight, not more than 2 % of its nominal breaking force — is applied, then one grip rotates at uniform speed in one direction, within a rate limit that depends on diameter. The result is the number of complete turns to fracture, or confirmation that the required count was reached, plus the fracture appearance. The required count is not in IS 1717; it comes from the product standard for the wire.

ASTM E143 — Shear modulus at room temperature. The specimen is a solid round or thin-walled tube of uniform section with a marked gauge length. Torque comes from the machine transducer, but the angle of twist is measured between two points on the specimen with a twist-meter (troptometer), not from grip rotation, so grip slip and machine compliance are excluded. Torque is applied slowly within the elastic range and G is taken from the slope of the straight portion of the torque–angle plot.

ETS torsion testing machines are supplied for testing to ASTM E143 and IS 1717. If your specification names a method from the "related" rows, send it to us before ordering so grips, speed range and report format can be checked.

4. Specimens, gripping and machine anatomy

Torsion machines see round, hexagonal and square bars, wire from below 1 mm to 8 mm, tubes, bolts and screws, shafts, springs, drill bits, drive shafts and welded shaft-to-flange joints. Round and hex sections go in three-jaw self-centring chucks, square bar in four-jaw or square collets; wire needs serrated grips with adjustable clamping so it is held without a stress-raising bite at the jaw edge; fasteners and shaft ends take hexagon or square sockets and threaded adaptors; tubes need an internal mandrel so the wall does not collapse; welded joints, springs and drive parts get fixtures machined to the drawing.

The machine itself is simple. A rigid bed carries two heads on a common axis. The drive head holds a servo motor and a planetary or worm gearbox that rotate the specimen in either direction at controlled speed, with an angle encoder on the spindle. The fixed head carries the second chuck on a strain-gauge reaction torque cell, so the reading is the torque in the specimen, not the motor output. It slides along the bed to set the gauge length and, on better machines, floats axially so a bar that shortens as it twists is not pulled into tension. A controller shows torque, angle and speed with settable limits, PC software records the curve, and guards, interlock and emergency stop complete the assembly.

5. Choosing the torque (Nm) capacity

Estimate the breaking torque of your largest specimen from its tensile strength, then apply the 20–80 % rule familiar from UTM capacity. Torsional shear strength of steel is roughly 0.6–0.75 × UTS, and a ductile solid round carries up to 4/3 of its first-yield torque before breaking. The upper figure and the plastic factor together give a convenient upper bound:

T_break (N·mm) ≈ τ_u × π d³ / 16 × 4/3   ≈   UTS (MPa) × π d³ / 16   (for τ_u ≈ 0.75 UTS)

Use the highest UTS you expect for the grade, the largest diameter you will grip, and the thread minor diameter for fasteners. The table applies the rule and lists the ETS class that keeps the test between 20 and 80 % of full scale.

Assumptions: solid round section, shear strength 0.75 × UTS, full plastic redistribution (4/3). Work-hardening grades can exceed these figures; brittle materials fall short. The 2 mm wire row shows why wire and modulus work rarely share a torque cell: 1.4 N·m is below the verified range of a 60 N·m transducer, acceptable for a twist count but not for a reported torque value.

Move up a class if the estimate lands above 80 % — a harder lot will one day trip the overload. If your work spans more than about a 10:1 torque ratio, specify a second, lower-range torque cell rather than one large frame.

6. Speed, accuracy, verification, reporting and safety

1. Speed. Modulus tests run slowly, a few °/min, so the two-degree elastic window yields hundreds of data points. Bar tests to fracture run at moderate speed. Wire twist counts run fast — tens of rpm — within the diameter-dependent limit in IS 1717, so a 100 d test piece finishes in under a minute. Specify a range that covers both ends, with closed-loop speed hold under load.
2. Torque accuracy. ±1 % of reading (class 1) over the range you use, normally from 10 or 20 % of capacity to full scale, in both directions if you test clockwise and anticlockwise.
3. Angle resolution. 0.1° is adequate for twist counts and angle to fracture; 0.01° or better, measured on the specimen, is needed for shear modulus to ASTM E143. Motor step counting includes gearbox backlash and wind-up and is no substitute for an encoder.
4. Verification of torque. Either a reference torque transducer with a traceable certificate is mounted in line with the specimen axis and compared with the machine reading at five or more points (ASTM E2624 describes the practice), or a calibration arm of known length is fitted to the fixed head and dead weights hung from it, giving T = m × g × L with the arm horizontal at each point. Repeat annually and after relocation, repair or torque-cell replacement.
5. Verification of angle. Compare the encoder against a graduated disc or precision index on the spindle, or an autocollimator, at several angles through a full turn.
6. Data and reporting. Torque–angle curve, the yield method used (offset angle or proportional limit), the regression window for G, maximum torque, angle or turns at fracture and fracture description, with specimen identity, dimensions, gauge length, speed, temperature, operator, machine identity and last calibration date, exported to CSV and PDF.
7. Safety. Specimens fail suddenly and wire ends whip outward. Keep a full interlocked guard over both grips, set torque and angle limits before every test, never reach in while the spindle moves, and wear eye protection for wire.

Our Extension Calibrator verifies displacement and extensometers to ISO 9513 / ASTM E83. It is not a torque calibration device; torque is verified with a reference transducer or calibration arm as above.

Frequently asked questions

Which ETS machine

• ETS-TTM-6(E), 60 N·mWire to IS 1717, spring wire, small bars to 6 mm and small springs.
• ETS-TTM-10(E), 100 N·m and ETS-TTM-20(E), 200 N·mMild steel bars to 8–10 mm, aluminium to 12 mm, small fasteners and shafts.
• ETS-TTM-50(E), 500 N·mThe general-purpose laboratory class: 10–12 mm alloy steel, 12 mm mild steel, M12 bolts, ASTM E143 modulus specimens.
• ETS-TTM-100(E), 1000 N·m and ETS-TTM-200(E), 2000 N·m16–20 mm bars, drive shafts, welded shaft joints and heavy fasteners.
• All torsion testing machinesDigital torsion testers from 60 to 2000 N·m with servo drive, reaction torque cell, angle encoder and in-house controller and software.