A Tensile Testing Machine is a measuring instrument that pulls a specimen apart along its axis while recording force and elongation in real time. Metallurgical labs and aerospace QA departments rely on it to certify every batch of bar stock, sheet, and forging that goes into airframes, pressure vessels, and fasteners. It works by gripping the sample between two crossheads, driving them apart at a controlled rate, and logging load-cell output against extensometer travel. The result is a stress-strain curve that yields modulus, yield strength, ultimate tensile strength, and elongation at break — the four numbers that decide whether the material ships or scraps.
Tensile Testing Machine Interactive Calculator
Vary specimen size, peak load, and elongation to see ultimate tensile strength, area, and engineering strain update on a tensile test diagram.
Equation Used
The tensile test converts peak load from the load cell into engineering ultimate tensile strength by dividing by the original specimen area. For a rectangular sample, A0 is width times thickness. Engineering strain is elongation divided by original gauge length.
- Engineering stress uses the original cross-sectional area.
- Rectangular specimen cross-section is assumed.
- Force is axial and evenly distributed through the gauge section.
- Strain is based on extensometer elongation over original gauge length.
How the Tensile Testing Machine Works
The machine has two crossheads. One is fixed, one moves. You clamp the specimen between matched grips on each crosshead, zero the load cell and the extensometer, then command the moving crosshead to pull at a defined rate — typically 5 mm/min for metals per ASTM E8, slower for elastomers, faster for some plastics. As the specimen stretches, the load cell on one crosshead reads the axial force in newtons, and the extensometer clipped to the gauge length reads elongation in micrometres. The control software plots load against extension, then converts to engineering stress (force divided by original cross-section) and engineering strain (elongation divided by original gauge length). That gives you the stress strain curve.
Why two separate sensors? Because crosshead travel includes machine compliance, grip slip, and specimen end-effects that are not the material's true elongation. The extensometer reads only the gauge section, so Young's modulus calculated from crosshead travel will under-read by 10-30% on a stiff steel. Use the extensometer for the elastic region and the upper yield, then switch to crosshead travel for elongation at break once the extensometer is removed before fracture.
Tolerances matter. The load cell must be Class 0.5 or better per ISO 7500-1 if you are reporting yield strength to a customer spec. Grip alignment must hold the specimen within 0.15 mm of the load axis or you will see a premature shoulder failure instead of a clean gauge-length break — and the ultimate tensile strength reading will come in low by 3-8%. Crosshead speed drift above ±10% of nominal will shift the yield point on strain-rate-sensitive materials like austenitic stainless or any polymer. The most common failures we see in returned data are: grip slip on hardened round stock (fix: serrated wedge inserts), extensometer knife-edge migration (fix: rubber bands plus correct knife pressure), and load-cell zero drift after a hard fracture impact (fix: re-zero between every test, not every batch).
Key Components
- Load frame: The structural backbone — typically a two-column or four-column design rated for the maximum force, with stiffness expressed in kN/mm. A 100 kN frame should have frame stiffness above 150 kN/mm so that frame deflection contributes less than 1% to measured strain at full load.
- Load cell: A strain-gauge transducer that converts axial force to a millivolt signal. Accuracy class 0.5 (ISO 7500-1) means the cell reads within ±0.5% of indicated value from 1% to 100% of capacity. Always pick a cell where your test load lands between 10% and 90% of capacity — a 100 kN cell used at 2 kN will report noise.
- Crosshead drive: Either a ball-screw electromechanical drive or a servo-hydraulic actuator. Ball-screw machines hold crosshead speed to ±0.5% from 0.001 to 500 mm/min. Servo-hydraulic systems handle higher loads and dynamic tests but cost 2-3× more and need a hydraulic power unit.
- Grips: Wedge grips for round and flat stock, pneumatic grips for thin sheet, threaded button-head grips for round specimens above 25 mm diameter. Grip face hardness must exceed specimen hardness by 10 HRC or you will mark the grip faces and bias the next test.
- Extensometer: A clip-on or non-contact device that measures elongation across a defined gauge length, typically 25 mm or 50 mm. Class B-1 per ASTM E83 resolves 1 µm. Required for any modulus or 0.2% offset yield reading — crosshead travel alone is not accurate enough.
- Control and data system: Closed-loop controller running at 1 kHz minimum sample rate, with software that captures the stress strain curve and computes yield strength, ultimate tensile strength, elongation at break, and reduction of area. Sample rate below 100 Hz will miss the upper yield peak on a low-carbon steel.
Where the Tensile Testing Machine Is Used
Anywhere a material claim has a number attached to it, a tensile testing machine generated that number. The instrument is the bridge between metallurgy on paper and certified material on a shop floor. You will find them in steel mills running heat-by-heat QC, in aerospace primes qualifying every coil of titanium sheet, in medical-device labs proving suture strength, and in university labs teaching first-year materials science.
- Steel production: ArcelorMittal mill labs running ASTM E8 round-bar tensile tests on every heat of rebar to certify yield strength above 420 MPa for Grade 60 deformed bar.
- Aerospace QA: Boeing supplier audits using Instron 5985 250 kN floor-standing UTMs to qualify Ti-6Al-4V forgings for 747 landing gear, with ultimate tensile strength specified at 895 MPa minimum.
- Medical devices: Ethicon testing Vicryl absorbable suture on a Zwick Roell zwickiLine 5 kN tabletop frame to certify knot-pull strength per USP <881>.
- Automotive fasteners: Nucor Fastener pulling Grade 8 bolts on a 600 kN Tinius Olsen Super L frame to verify proof load and tensile strength per SAE J429.
- Polymer R&D: Dow Chemical running ISO 527 dogbone tests on new HDPE pipe-grade resins at 50 mm/min to map yield stress and elongation at break.
- Composites: Hexcel testing carbon-fibre prepreg coupons per ASTM D3039 on an MTS Criterion 45 to certify modulus above 135 GPa for wing-skin lay-ups.
- Civil engineering: ICC-ES accredited labs pulling concrete reinforcement anchors per ASTM E488 to qualify post-installed adhesive anchors for seismic zones.
The Formula Behind the Tensile Testing Machine
The core calculation is engineering stress — force over original cross-sectional area. The number you care about is ultimate tensile strength, σUTS, the highest engineering stress the specimen reaches before necking. At the low end of the typical operating range, a soft 1018 mild steel rod gives you around 440 MPa and a long ductile plateau. A nominal 4140 alloy steel comes in around 850 MPa with a sharp yield. Push to the high end with a quenched-and-tempered 4340 and you'll see 1,250 MPa plus, but the elongation at break collapses from 25% on the mild steel to 10% on the 4340 — you trade ductility for strength along the entire range.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| σUTS | Ultimate tensile strength (peak engineering stress) | MPa (N/mm²) | psi or ksi |
| Fmax | Maximum axial load recorded by the load cell during the test | N | lbf |
| A0 | Original cross-sectional area of the gauge section before loading | mm² | in² |
Worked Example: Tensile Testing Machine in a wind-turbine bolt manufacturer in Esbjerg
A wind-turbine fastener manufacturer in Esbjerg is qualifying M42 × 3 hot-dip galvanised tower bolts to ISO 898-1 Property Class 10.9. The tensile test specimen is a machined round per ISO 6892-1, gauge diameter 12.0 mm, gauge length 60 mm. They run the test at 10 MPa/s stress rate on a 600 kN Zwick Roell Z600E. Property Class 10.9 requires ultimate tensile strength between 1,040 and 1,220 MPa.
Given
- d0 = 12.0 mm
- Fmax,nominal = 127,000 N
- L0 = 60 mm
Solution
Step 1 — compute the original cross-sectional area from the gauge diameter:
Step 2 — at the nominal recorded peak load of 127,000 N, the engineering ultimate tensile strength is:
That sits cleanly in the middle of the Class 10.9 window — exactly where a healthy heat should land. Step 3 — at the low end of the spec window, a borderline-pass batch reading 117,600 N gives:
That is the absolute floor. Anything below this and the bolt fails the property class — the QA department holds the lot. At the high end, a heat that came in hard reads 138,000 N:
Right at the ceiling. Bolts above 1,220 MPa have reduced toughness and are vulnerable to hydrogen embrittlement after galvanising — the Esbjerg lab will reject those just as quickly as low ones, because a brittle tower bolt at -20 °C in the North Sea is a service failure waiting to happen.
Result
The nominal ultimate tensile strength is 1,123 MPa — a clean Class 10. 9 pass. In practice the operator sees the load curve climb steeply through elastic, roll over at about 102 kN (yield), keep rising to 127 kN over another 4-5% strain, then drop sharply as necking starts in the gauge centre. Across the operating window, 1,040 MPa is the rejection floor and 1,220 MPa is the ceiling, with the sweet spot sitting at 1,100-1,150 MPa where strength and toughness both meet the spec. If your measured σUTS reads 5-10% low, suspect specimen taper from a worn lathe tool — a 0.05 mm taper across the 60 mm gauge length shifts the calculated stress because failure happens at the smallest cross-section, not the measured one. If the curve shows a stepped or ragged plastic region, the wedge grips are slipping on the galvanised threads — switch to threaded button-head ends. If the test reports a low yield but normal UTS, the extensometer knife edges have migrated on the slick zinc surface and missed the true elastic region.
Choosing the Tensile Testing Machine: Pros and Cons
A tensile testing machine is the gold standard for material certification, but it is destructive, slow, and not the right tool for every job. Two alternatives come up constantly: portable hardness testers, which give an indirect strength estimate non-destructively, and impact testers like Charpy, which characterise toughness rather than tensile strength. Pick the wrong instrument and you either over-test cheap material or under-test critical material.
| Property | Tensile Testing Machine (UTM) | Portable Hardness Tester (Leeb/Rockwell) | Charpy Impact Tester |
|---|---|---|---|
| Measured property | Yield, UTS, elongation, modulus directly | Hardness only — strength inferred via correlation | Impact toughness in J, not strength |
| Accuracy on UTS | ±0.5% with Class 0.5 load cell | ±10-15% via hardness-to-strength conversion | Not applicable |
| Test time per specimen | 3-15 minutes plus machining | 30 seconds, no specimen prep | 1-2 minutes per pre-notched specimen |
| Specimen destruction | Yes — specimen is pulled to failure | No — small indent only | Yes — specimen is broken |
| Capital cost | $25k tabletop to $250k floor-standing | $1.5k-$8k portable | $15k-$60k pendulum |
| Best application fit | Material certification, R&D, QC of every heat | Field hardness checks, weld verification | Low-temperature toughness for pressure vessel and ship steel |
| Crosshead speed range | 0.001-500 mm/min electromechanical | Not applicable | Fixed pendulum velocity ~5 m/s |
Frequently Asked Questions About Tensile Testing Machine
That drop is the upper and lower yield point — it is real, not an artefact, and it appears on low-carbon steels and some annealed mild steels because dislocations are pinned by interstitial carbon and nitrogen until they break free in a sudden burst (Lüders bands). You will see it sharply on 1018 hot-rolled but barely on cold-drawn versions of the same alloy.
If the drop looks jagged or stepped instead of clean, your data sample rate is too low — push it above 500 Hz, ideally 1 kHz, so the controller actually captures the peak before strain accommodates.
It depends on what your customer spec calls out. ASTM A370 for carbon steel allows upper yield. ASTM E8 for general metals defaults to 0.2% offset. ISO 6892-1 reports both ReH (upper) and Rp0.2 (offset). The two values can differ by 20-40 MPa on a sharp-yielding mild steel, so picking the wrong one can move a borderline material from pass to fail.
Rule of thumb: if the curve has a clear yield point drop, report upper yield. If it's a smooth curve like aluminium or stainless, you must use 0.2% offset — there is no other defensible yield definition.
Pick the cell so your expected peak load lands between 10% and 90% of cell capacity. Below 10% the cell signal-to-noise ratio degrades — a 100 kN cell used to break 2 kN suture has roughly 50× more noise than the actual signal. Above 90% you risk overload damage on the next stiff specimen.
For mixed-material labs, buy two cells: a 5 kN for polymers, sutures, and thin sheet, and a 100-250 kN for metals and fasteners. Frame manufacturers like Instron and Zwick let you swap cells in under five minutes with auto-recognition.
Almost always one cause: you are calculating modulus from crosshead travel instead of from the extensometer. Crosshead travel includes load-frame compliance, grip slip, and specimen end-effects, all of which add apparent strain that is not happening in the gauge length. On a stiff steel at 200 GPa, this typically reads as 130-160 GPa from crosshead data.
Fix it by clipping a Class B-1 extensometer across the gauge length and computing modulus only from extensometer strain in the linear region between 10% and 50% of yield. Remove the extensometer before fracture if it's a contacting type — a steel knife-edge does not survive a snap-back from a 600 MPa break.
4% scatter on a single heat is too much. Three things cause it in order of frequency: specimen machining inconsistency (gauge diameter varying ±0.05 mm batch to batch — measure every specimen with a calibrated micrometre at three points and average), grip misalignment introducing a bending component (check concentricity of upper and lower grips with a dial indicator — must be within 0.15 mm TIR), and crosshead speed drift on older ball-screw machines.
Tighten specimen prep first. Most labs blame the machine when 80% of scatter is sitting in the lathe operation that made the specimens.
Yes, and for fasteners it is often required. ISO 898-1 and ASTM F606 both define full-size bolt tensile tests where you grip the head and a threaded fixture pulls on the nut end. The reported strength uses the stress area As calculated from the thread pitch and basic diameter, not the major diameter.
The catch: full-size tests measure the weakest section, which is usually the thread root, so the result is a system strength not a material strength. If you need the material UTS independent of the thread geometry, you must machine a separate round specimen from the bolt shank.
References & Further Reading
- Wikipedia contributors. Universal testing machine. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
