A tensile test looks simple from across the laboratory. A specimen goes into the grips, the crosshead moves, the specimen stretches and breaks, and the software prints a few numbers. Most of what decides whether those numbers are right happens before the first newton of force: how the specimen was machined and measured, where the gauge marks sit, how the specimen hangs in the grips, which extensometer is on it and how fast the machine is told to pull.
This guide walks through a room-temperature tensile test on metal in the order a laboratory performs it, following the two standards most reports are written against, ASTM E8/E8M and ISO 6892-1. Where the two differ, it says so, because a result is only comparable with another result obtained under the same rules.
What a Tensile Test Measures
The principle is the same in both standards: a test piece is strained by a tensile force, generally until it fractures, and one or more properties are calculated from the force and the extension. ASTM E8/E8M lists them as yield strength, yield point elongation, tensile strength, elongation and reduction of area, for metallic materials in any form. ISO 6892-1 defines the same properties with its own symbols, and every stress in it is an engineering stress: force divided by the original cross-sectional area, So.
Room temperature is not quite the same range in the two documents. ISO 6892-1 tests between 10 °C and 35 °C unless something else is specified, with 23 °C ± 5 °C where the conditions are controlled, and asks for the temperature to be recorded when the test falls outside that window. ASTM E8/E8M treats 10 °C to 38 °C as room temperature. Tests at higher temperatures move to other standards, ISO 6892-2 and ASTM E21.
The properties fall into two groups. Strengths are stresses: the yield or proof strength where plastic deformation begins, and the tensile strength, Rm, at the maximum force. Ductility is reported as elongation after fracture, A, and reduction of area, Z, both measured on the broken specimen. ASTM adds a caution worth keeping in mind for both groups: a specimen machined from a selected part of a product may not represent the whole product or how it behaves in service.
Before the Test: the Specimen
ISO 6892-1 allows circular, square, rectangular, annular and other uniform cross-sections. Machined test pieces have a parallel length joined to the gripped ends by a transition radius, and the parallel length always has to be longer than the original gauge length, Lo. Products with a uniform section, such as bars, wires and sections, can be tested without machining, provided the gauge marks stay a reasonable distance from the grips.
The gauge length itself follows a rule. For proportional test pieces, ISO 6892-1 sets Lo = 5.65 √So, which on a round bar comes to five diameters, and Lo may not be shorter than 15 mm. ASTM differs between its two versions: for most round specimens the gauge length is four diameters in E8 and five diameters in E8M, which ASTM itself calls the most significant difference between the two. On the standard 12.5 mm round that means 50 mm under E8 and 62.5 mm under E8M. The difference matters at the end of the test, because elongation measured over four diameters is not comparable with elongation over five.
Machining decides more than the shape. The surface of the reduced section should be free of notches and burrs, and the edges of flat specimens are machined rather than sheared. Then the specimen is measured, using its actual dimensions rather than the nominal ones. The arithmetic is unforgiving: an area measured 1 % too large lowers every stress result on the report by 1 %. ASTM asks for measuring devices traceable to national standards and for reasonable effort to keep errors from measurement, resolution and rounding below 1 %.
Marking the Gauge Length
Elongation after fracture is measured between gauge marks, so the marks are part of the measurement. They have to be fine marks or light scribed lines: a mark pressed too deep is a surface defect in the reduced section and can decide where the specimen breaks. Many laboratories mark the whole reduced section at an even pitch rather than placing only two end marks. That costs nothing at the start and pays off when a specimen breaks off-centre, because ISO 6892-1 has an informative annex that, by agreement, re-evaluates the elongation by counting equal intervals either side of the fracture.
Our article on gauge length marking and the elongation result covers proportional and fixed gauge lengths and the interval counting in detail. Automatic marking takes the pitch out of the operator’s hands: the GenMark 500 marks round and flat specimens at a switchable 5 or 10 mm pitch, to within 0.05 mm or 0.5 %, whichever is greater.
Gripping and Aligning the Specimen
ISO 6892-1 states the requirement in one line: the axis of the test piece shall coincide with the axis along which the force is applied. Everything about gripping serves that line. ASTM E8/E8M gives more detailed guidance on wedge grips, threaded and shouldered specimens, sheet and wire, and notes that any departure of the specimen axis from the centre line of the machine heads can introduce bending.
The grips are rarely neutral. A review of tensile testing by the UK National Physical Laboratory brought together measurements made with strain-gauged round specimens: alignment was best with threaded grips, worse with button-head collet grips and worst with self-tightening wedge grips, where the bending stress at yield was still more than 30 % of the direct stress. The cause was tiny differences in how far the serrations bit into each side of the specimen. The same review notes that universal joints, relied on to let the load train align itself, tend to lock as the load rises. Wedge grips remain the standard choice for most metal specimens, which is a reason to check alignment rather than a reason to avoid them. Our guide to UTM grips and fixtures compares the types.
Bending matters because it changes the numbers. Misalignment lowers the apparent proof or yield strength, can suppress the upper yield point and falsifies any modulus taken from the curve. ASTM E1012 is the practice for measuring it: strain gauges on a specimen show how much bending the load train adds at a given force. E1012 itself sets no pass limit. The acceptable percentage comes from the test method that refers to it, or from the customer, and the VAMAS procedure behind the international alignment work recommends checking at least after a force calibration, after any repair or reassembly of the load train, and after the crosshead has been moved, unless that move is shown to make no difference. Our Alignment Device sits between the crosshead and the grips, with coarse and fine adjustment and a built-in coaxiality meter, so a frame can be brought back onto its axis after each of those events.
Zero, Clamp, Then the Extensometer
The force zero is set with the grips assembled but before the specimen is clamped, as the ISO tensile standards describe it, so the weight of the grips is compensated and the clamping force never becomes part of the reading. The specimen is then clamped firmly enough that it cannot slip and re-seat while the curve is still elastic; slippage shows up as steps in the elastic line.
The extensometer goes on next. ISO 6892-1 makes it mandatory for proof strength, yield point extension and extension at maximum force: crosshead movement includes the compliance of the frame, grips and jaw seating and is not strain. For proof and yield strength the extensometer has to meet ISO 9513 class 1 or better, and a class 2 device is accepted for properties measured at larger extensions. ASTM E8/E8M asks for class B-2 or better under ASTM E83 for yield. Our guide to extensometer classes under ASTM E83 and ISO 9513 explains the two systems. A clip-on extensometer is usually removed after yield to protect it, and elongation after fracture is then taken from the gauge marks.
Running the Test at the Right Speed
Speed is the setting most often treated as a detail, and the one with the clearest effect on yield. Studies collected by NPL found the lower yield strength of carbon steels rising by about 10 MPa for every tenfold increase in strain rate, measured on dozens of steels in several countries. On one steel tested on nine machines the sensitivity was about 10 MPa per decade with a scatter of around ± 19 MPa, and it fell to 6.6 MPa per decade when the strain rate was held under tight extensometer control.
Holding a constant crosshead speed does not hold a constant strain rate. Before yield, much of the crosshead movement goes into stretching the load train, so the specimen strains slowly; once it yields, nearly all of the movement goes into the specimen. One study measured the specimen’s strain rate changing by a factor of about 4.8 between the elastic region and the region after yield at the same crosshead speed.
ISO 6892-1 handles this with two methods. Method A is based on strain rates and is strongly recommended, because it keeps the rate steady while rate-sensitive properties are measured. It works either in closed loop from the extensometer, A1, or at a constant crosshead separation rate calculated from the parallel length, A2. Its four ranges are 0.00007, 0.00025, 0.002 and 0.0067 per second, each within ± 20 %. Upper yield, proof and total-extension strengths are measured in range 1 or 2; lower yield strength and yield point extension in range 2 or 3; tensile strength, elongation and reduction of area in ranges 2 to 4. If only the tensile strength is needed, range 3 or 4 may be used throughout. Method B sets a stress rate in the elastic range instead: 6 to 60 MPa per second for metals with a modulus of 150 GPa and above, such as steel, and 2 to 20 MPa per second below that, as for aluminium alloys, brass and titanium.
ASTM E8/E8M offers three ways to control the test through yield: a stress rate of 1.15 to 11.5 MPa per second in the elastic range, a closed-loop strain rate of 0.015 ± 0.006 per minute from the extensometer, or a crosshead speed of 0.015 ± 0.003 per minute of the reduced section length. After yield, the rate rises to 0.05 to 0.5 per minute of the reduced section until fracture, unless the material specification says otherwise. The method used is part of the report. In practice this means the machine has to change speed partway through the test. Rates expressed per minute convert simply to crosshead speeds: range 2 on a 75 mm parallel length is about 1.1 mm per minute, and range 4 about 30 mm per minute.
That is why modern tests run from a stored method rather than a speed knob. GenTest, the software on NextGen machines, builds each test from steps with changeover conditions, so the switch from the yield phase to the post-yield phase happens at the same point on every specimen, and it calculates ReH, ReL, Rp, tensile strength and elongation as the method defines them. Our overview of universal testing machine software covers what to look for.
After Fracture: Reading and Reporting the Result
The strengths come straight from the record. Where the material shows a yield point, ReH is the highest stress before the first drop in force and ReL the lowest stress during yielding, ignoring the initial transient. Where it does not, the proof strength Rp0,2 is read where a line parallel to the elastic part, offset by 0.2 % plastic extension, crosses the curve. ISO 6892-1 notes that the slope used for that line need not equal the modulus. The tensile strength Rm is the stress at the maximum force.
The ductility results need the specimen. The two broken halves are fitted carefully back together so that their axes lie in a straight line, the final gauge length Lu is measured, and the elongation after fracture is A = (Lu − Lo) / Lo × 100. Reduction of area compares the smallest cross-section after fracture, Su, with the original: Z = (So − Su) / So × 100. Both standards also limit how close to a gauge mark the fracture may be for the elongation to stand, which is why the evenly spaced marks described earlier are worth having.
These two results are the least reproducible part of the test. In round-robin programmes reviewed by NPL, the spread of elongation after fracture between laboratories was roughly 6 % to 28 % of the mean, against roughly 1.3 % to 5 % for tensile strength, and reduction of area depended strongly on the operator measuring the neck. Careful fitting, a sharp measurement and a consistent gauge length are what narrow that spread.
The report then states what was done. ISO 6892-1 reports strengths to the nearest whole MPa, yield point extension to 0.1 %, other extensions and elongations to 0.5 % and reduction of area to 1 %, and asks for the control mode and testing rate to be stated where they differ from its recommended methods. ASTM E8/E8M asks for the specimen type, the yield strength and the method used to determine it, yield point elongation, tensile strength and elongation, with reduction of area and uniform elongation where required. Whatever the standard, the gauge length belongs next to the elongation.
Keeping the Machine Verified
The force reading is verified against standards that cover the whole machine. ISO 6892 tensile testing calls for a force-measuring system of class 1 or better under ISO 7500-1, where class 1 allows a relative error of ± 1.0 % and class 0.5 allows ± 0.5 %. ASTM E4 has no classes and verifies the machine to ± 1 %. Our comparison of ISO 7500-1 and ASTM E4 sets the two side by side. Extensometers are calibrated separately, under ISO 9513 or ASTM E83, and alignment is checked under ASTM E1012 as described above.
Verification also has a practical side that an audit will look at: grip faces that are worn or clogged, an extensometer that slips, a load cell that drifts between calibrations. Our article on why material test results fail audits goes through the equipment causes laboratories tend to miss.
- The area measured before the test sits under every stress: 1 % too large lowers every strength by 1 %.
- Gauge length travels with elongation: four diameters in ASTM E8, five in E8M and ISO 6892-1 on most rounds.
- Misalignment bends the specimen and lowers the apparent proof or yield strength, so alignment is checked after a force calibration, a repair or a crosshead move.
- Elongation is the least reproducible result: roughly 6 % to 28 % of the mean between laboratories, against 1.3 % to 5 % for tensile strength.
- The force reading is verified to class 1 or better under ISO 7500-1, or to ± 1 % under ASTM E4.
Equipment for the Whole Workflow
A frame for metals tensile testing has to cover the full range of speeds the standards ask for, from the slow elastic and yield phases to the faster post-yield phase. The NextGen NG-EML electromechanical machines are calibrated to class 0.5 under ISO 7500-1 and run from 0.00005 mm per minute upward, which covers both ends of the ranges above. For high forces, such as large rounds, plate and reinforcing bar, servo-hydraulic frames take over; our comparison of electromechanical and servo-hydraulic machines covers the choice.
Keep Reading in Electromechanical Universal Testing Machines for Plastic Testing
See the Electromechanical Universal Testing Machines for Plastic Testing range
Flexural (Bend) Testing Explained: 3-Point vs 4-Point, Standards, and SetupFlexural (bend) testing explained: what it measures, three-point vs four-point setups, the standards for plastics, metals, composites and concrete, and how to choose the right fixtures and...Read the articleThe Rockwell Scale Explained: How to Choose Between HRC, HRB and Superficial ScalesA Rockwell hardness number tells you very little until you know the letters that come with it. A reading of 60 on one scale and a reading of 60 on another describe different materials, tested with a different indenter under a different force, and neither can be compared with the other until someone knows which scale produced it. A drawing that asks for 60 HRC is specifying a whole test: a diamond indenter, a total force of 150 kgf and a depth measurement made in a defined way. ASTM defines thirty Rockwell scales. Picking the right one is a decision that can go wrong without any warning, because the wrong scale does not produce an error message. It produces a number, and that number looks just as plausible as a valid one. This article explains what...Read the articleHow to Measure Case Depth: Effective Depth, Total Depth and the Hardness Limit That Decides ItA drawing asks for case depth in a single number, and that number carries a tolerance, an acceptance decision and sometimes a rejected heat lot. What the drawing rarely says is which of two different depths it means, or which hardness value marks the boundary. Two laboratories can follow the same procedure on the same part, measure carefully, and report depths that differ by a noticeable margin, because they drew the line in different places. The measurement itself is not the hard part. Agreeing on what is being measured is. This article is about the traverse: how case depth is defined, what fixes the boundary, and what the work demands from the equipment that produces it. If you are still deciding whether microhardness is the right method at all,...Read the articleVicat Testing and What Cement Setting Time Really Tells YouVicat setting time is a standardized cement paste measurement used for specification compliance and quality control. It identifies when a paste prepared under defined conditions reaches the initial and final set criteria of the selected method. It does not directly predict concrete haul time, cold joint formation, finishing time or field workability, because those depend on the complete mixture: water content, admixtures, aggregates and temperature. This guide covers what Vicat setting time measures, how the apparatus produces the number, which test conditions make results drift, and what separates a manual instrument from an automatic one. Setting Time Is Not Early Strength Initial and final set describe the stiffening of a standardized cement...Read the articleHow to Choose the Right Universal Testing Machine for Plastic TestingChoosing a universal testing machine for plastic testing often sounds easier than it really is. Many buyers start with force capacity, frame size, or the idea of buying extra room for future work. That may seem like a safe decision, but it can push a lab toward a system that costs more, takes up more space, and does not match daily testing needs. The challenge comes from how broad plastics testing can be. One lab may focus on tensile testing for rigid molded specimens, another may run flexural tests on semi-rigid components, while others may handle thin films, sheets, reinforced plastics, or a wider mix of quality control and development work. Once applications vary, the machine can no longer be chosen by force range alone. A better decision starts...Read the articleMachines Behind This Article
Electromechanical Universal Testing Machines for Plastic TestingNG-EML Series A – Single-Column Universal Testing Machine (50 N – 5 kN)NG-EML Series A is a compact single-column benchtop universal testing machine for low-force tensile, compression, and flexural testing. Covering 50 N to 5 kN...View product
Electromechanical Universal Testing Machines for Plastic TestingNG-EML Series B – Dual-Column Universal Testing Machine (100 N – 10 kN)NG-EML Series B is a dual-column benchtop electromechanical universal testing machine for precision testing from 100 N to 10 kN. Built for research and industrial...View product
Electromechanical Universal Testing Machines for Plastic TestingNG-EML Series C – Dual-Column Universal Testing Machine (5–50 kN)NG-EML Series C is a dual-column electromechanical universal testing machine available in bench-top and floor-standing formats from 5 kN to 50 kN. Built for...View product
Electromechanical Universal Testing Machines for Plastic TestingNG-EML Series D – Floor-Standing Universal Testing Machine (50–1000 kN)NG-EML Series D is a heavy-duty floor-standing electromechanical universal testing machine for high-capacity tensile, compression, and flexural testing. Available...View product
ServoNG-SHM Class D – Servo-Hydraulic Universal Testing MachineNG-SHM Class D is a servo-hydraulic universal testing machine for high-force mechanical testing of metals and structural components. Available in 600 kN, 1000 kN,...View product
Universal Testing Machines for Metal TestingGenMark 500 – Automatic Gauge Length Marking MachineThe GenMark 500 is a full-automatic gauge length marking machine for metal bar, rebar and wire rod, applying the reference marks used to determine elongation after...View product
Electromechanical Universal Testing MachinesWedge Tensile Grips for Metals and CompositesWedge tensile grips are built for high-load tensile testing of flat and round specimens made from metals, composites, and plastics. Their wedge-action mechanism...View product
Electromechanical Universal Testing MachinesDual-Action Hydraulic Wedge Grips for High-Capacity TestingDual-Action Hydraulic Grips are high-capacity wedge-clamping fixtures for tensile testing on universal testing machines. Designed for metals, plastics, and...View product
Universal Testing Machines for Metal TestingAlignment Device – NADCAP-Ready Precision FixtureThe Alignment Device is a NADCAP-ready precision fixture for verifying and correcting load-frame alignment on universal and fatigue testing machines. Designed to...View product
Electromechanical Universal Testing Machines for Plastic TestingGenTest – Advanced UTM Testing SoftwareGenTest is NextGen’s advanced UTM testing software for electromechanical universal testing machines, managing test setup, live control, data acquisition,...View product