A test request lands on the bench with a material, a specimen drawing and a list of values to report. Before the first specimen goes into the frame, someone has to decide whether to fit an extensometer. Mounting one takes setup time, and university lab procedures have operators take clip-on units off before failure to keep them from being damaged. Leaving the device off altogether can make one of the reported values impossible to defend.
The answer comes from the list of results, not from the material or the size of the machine. Tensile strength, reduction of area and elongation measured from gauge marks can all be produced without one. Proof strength, extension at maximum force and a usable modulus cannot, and the test method then also decides how accurate the device has to be.
The extensometer also depends on the parts that hold the specimen. Labs typically pick wedge, pneumatic and self-tightening grips by material and force, and the choice matters for strain data too. Grips that hold the specimen without slipping or crushing it, and extensometers chosen for the material and the method, sit side by side among the factors that move a tensile result.
What follows sorts the common tensile results by what they need, then covers the accuracy class and the type of device, and finishes with the questions labs raise most often before clipping one on.
Results That Rest On Force And Gauge Marks Alone
Several standard results can be produced without any extension reading. Tensile strength is the plainest case. ISO 6892-1 defines it as the stress corresponding to maximum force, and the value comes from the peak of the force record divided by the original cross-sectional area.
Upper yield strength is also defined through force. ISO 6892-1 describes it as the maximum stress before the first decrease in force, so the defining event is a drop in the force record rather than a strain value.
A round robin on pure iron, published as NIST Technical Note 1879, shows this in practice. One participant ran a set of tests without an extensometer and estimated offset strength from force and actuator data. The report adds that most of those yield values were lower yield strengths, which it describes as unaffected by whether actuator displacement or specimen elongation was used.
Elongation after fracture can be measured by hand. The broken halves are fitted back together so their axes lie in a straight line, the distance between gauge marks applied before the test is measured, and the increase is expressed as a percentage of the original gauge length. Reduction of area compares the smallest cross-section after fracture with the original one. Teaching labs work the same way: one university procedure has students remove the extensometer at the yield point and then follow elongation between punched gauge marks with dividers and a scale.
The gauge marks then become part of the measurement. Once the specimen breaks, nothing re-measures the original spacing, so an error made when the marks were applied passes straight into the reported elongation. ISO 6892-1 also gives percentage elongation after fracture and percentage total extension at fracture separate definitions and symbols, so a report needs to make clear which of the two it states.
Where The Report Cannot Be Written Without One
The line falls at any property defined by extension during loading. ISO 6892-1 says so directly in its definition of extensometer gauge length: for properties based wholly or partly on extension, such as proof strength, yield point extension and plastic extension at maximum force, the use of an extensometer is mandatory.
ASTM E8 covers yield strength, yield point elongation, tensile strength, elongation and reduction of area, and for yield it asks for an extensometer of Class B-2 or better under ASTM E83. The reason is easy to see in a teaching lab. One university procedure explains that actuator displacement includes the elastic deformation of the machine, that the percentage error from it is large early in the test while the specimen deformation is small, and that an extensometer is used in that initial part for a direct measurement in the gauge region. Students there calculate offset yield from both extensometer data and grip data, for comparison.
The NIST round robin shows what happens to a strain-based value measured the wrong way. Participants also reported uniform elongation, but three of the four laboratories had estimated it from actuator displacement rather than specimen elongation, and the parameter was left out of the analysis.
Strain Rate Control Can Require One Before Any Result Does
Some test plans need an extensometer for control, even when no strain-based value appears on the report. ISO 6892-1 Method A works from strain rate and can run in two ways: in closed loop from the extensometer, known as A1, or at a constant crosshead separation rate calculated from the parallel length, known as A2. Choosing A1 means the device has to be on the specimen whenever the loop is closed on it.
ASTM E8 also lists a closed-loop strain rate from the extensometer as one of three ways to control the test through yield, alongside a stress rate in the elastic range and a crosshead speed. After yield the rate rises until fracture, so in practice the machine has to change speed partway through the test.
Running Method A1 calls for testing frames whose controllers switch between load, displacement and strain control within a single test, for example moving from load control to strain control at the yield point. The slowest of the four ISO 6892-1 strain rate ranges is 0.00007 per second, so the controller also has to hold a very low rate steadily in strain control.
Modulus Asks For More Than A Yield Test
Reporting the elastic slope raises the bar again. ISO 6892-1 points out that the slope of the elastic part of the stress-extension curve may not represent the modulus of elasticity, and it describes the modulus calculation in a separate annex that is not needed for proof strength.
Work by the UK National Physical Laboratory explains why. Its report on modulus measurement in the tensile test concluded that good modulus data can be obtained, but that this generally requires a separate and dedicated setup with high-quality averaging strain measurement focused on the early part of the curve. It describes these as specialised tests that may not be realistic to run cost-effectively on a high-throughput machine.
The same report puts numbers on the problem. A Class 1 extensometer carries a total bias error of 1 percent or 3 micrometers, whichever is greater, and on a 25 mm gauge length at 0.1 percent strain the resulting error in modulus can be as high as 12 percent. For the highest accuracy it recommends a Class 0.2 averaging, high-resolution extensometer calibrated over the restricted strain range of the test, and it notes that strain gauges bonded to each side of the specimen are an alternative worth considering.
ASTM handles modulus in a method of its own. ASTM E111 covers Young’s modulus, tangent modulus and chord modulus, and among the variables it lists for precise determination are alignment and the ratio of error in extension measurement to the range of extension values used.
Plastics follow similar logic under their own methods. ISO 527-1 expects strain to be measured over a defined gauge length when modulus and accurate strain values are required. Laboratories working to ISO 527-2 see the practical effect: the same frame can report different values depending on whether strain comes from crosshead travel or from a suitable extensometer.
Accuracy Class Follows The Property And The Strain Range
ASTM E83 classifies extensometer systems by the magnitude of their errors into Classes A, B-1, B-2, C, D and E. ISO 9513 grades them into classes that include 0.2, 0.5, 1 and 2. A tensile method specifies the class each result needs, so the class a lab needs is set by the most demanding value in its test plan.
For metals the main requirements and recommendations line up as follows:
- Yield to ASTM E8: Class B-2 or better under ASTM E83.
- Proof strength to ISO 6892-1: class 1 or better under ISO 9513, with class 2 accepted for properties measured at larger extensions.
- Modulus: a Class 0.2 averaging extensometer is the National Physical Laboratory recommendation for the highest accuracy.
The class is tied to a range as well. Calibration results can depend on how the extensometer is mounted and aligned, on the gauge length used for the test program and on whether the calibrated range matches the strains actually used for reporting or for closed-loop control. Verification may also be needed after changes to the sensor, electronics, cabling, software scaling or contact parts.
Clip-On Or Non-Contact Depends On What Happens At Break
ASTM E83 also sorts systems into three types by how gauge length is handled. A Type 1 system, such as a clip-on extensometer, both defines gauge length and senses extension. A Type 2 system senses extension over a gauge length defined by features of the specimen. A Type 3 system senses strain ratiometrically, as a video camera system does. ISO 9513 covers axial and diametral systems, contacting and non-contacting, so a video extensometer can be calibrated and classed under the same standard as a contact one.
The deciding question is usually whether the device can stay on until fracture. University procedures treat clip-on units as fragile and expensive. One has students remove the extensometer when strain reaches 0.1 or the specimen starts to neck. Another has the test software prompt removal at a preset strain before failure, to avoid damaging the device, and the test then continues to fracture. The NIST round robin records labs removing theirs at roughly 0.7 mm, 1 mm and 5 to 6 mm of elongation. Some commercial test software can likewise prompt the operator when it is time to remove the device.
What happens after removal has to be planned. One university lab follows the rest of the test from actuator position, accepting that machine deformation is mixed into the reading. Others read elongation from gauge marks. Either way, the values that depend on accurate strain have to be captured before the device comes off.
Polymers and elastomers shift the balance toward travel. Plastics tensile work uses contact or non-contact extensometers depending on the strain range and the specimen geometry. Elastomers stretch far enough that rubber test setups commonly need extensometry with long travel, with the requirement set by the rubber type and the expected elongation.
Questions Labs Ask Before Clipping One On
Four questions come up repeatedly when a lab writes its tensile procedures.
1. Can Elongation Be Reported Without An Extensometer?
Yes, as elongation after fracture, measured between gauge marks on the fitted halves. ISO 6892-1 defines that value separately from total extension at fracture, which is read during the test, so the report should say which one it gives.
2. Does An Order For Tensile Strength Alone Need One?
No. Tensile strength is the stress at maximum force, worked out from the peak force and the original cross-sectional area. An extensometer becomes necessary once the same order adds yield or proof strength, extension at maximum force or modulus.
3. Can One Extensometer Cover Both Yield And Elongation?
That depends on its travel and its class. A clip-on unit chosen for yield is often removed before fracture, after which elongation comes from gauge marks or from machine movement. A non-contact system, or a contact unit with long travel, can stay on longer, provided its class and calibrated range cover both the yield region and the strains reported later.
4. Do ASTM E83 And ISO 9513 Cover Video Extensometers?
Both do. ASTM E83 includes ratiometric systems such as video camera systems as its third type, and ISO 9513 covers non-contacting systems alongside contacting ones. A non-contact system is therefore classed against the same requirements and has to meet the class the test method asks for.
