Specimens, dimensions and test conditions

Cutting a specimen is part of the test

A selected piece has an origin, orientation, preparation history and dimensions. Cutting, polishing or machining can change its surface or introduce damage. Record these steps rather than assuming the piece remained identical to its parent material.

Imagine two strips labelled only “rolled sheet”. One was cut parallel to rolling near an edge; the other was cut across rolling from a different sheet and polished on one face. If they respond differently, their labels do not isolate a material-direction effect. Origin, direction and preparation all changed. A useful specimen identity connects the individual piece to its parent stock, position, face and sequence of operations.

Rolling can create a preferred arrangement of grains or elongated features. A property may then depend on direction: anisotropy. Mark a material direction before cutting and preserve its relation to the eventual load or measurement direction. Parallel and transverse are incomplete without saying parallel or transverse to what. The direction of an arrow on a test-machine photograph is not automatically the rolling direction.

Preparation can also change what is being tested. Removing a layer changes the surface and dimensions; mechanical treatment can change near-surface residual stress, meaning stress that remains without an external load. In a surface-sensitive measurement this may be part of the signal rather than a removable nuisance. A smooth-looking surface is not evidence that the original near-surface state survived. Record the treated face, operation and relevant checks; a preparation choice suitable for imaging is not automatically suitable for measuring an untouched surface.

Follow one coupon from sheet to gauge

The first figure is an original hypothetical cutting record for sheet S7, represented by a 100 mm × 80 mm rectangle. Its +x direction is the marked rolling direction; +y lies across it. C1 is a nominal 60 mm × 6 mm strip parallel to +x. C2 has the same nominal plan dimensions but its length follows +y. The two marked regions do not overlap. The map specifies origins and orientations; it neither shows microstructure nor establishes that those two locations are representative of the whole sheet.

For C1, keep a sequence such as parent S7 → mapped C1 blank → edge finishing → measured coupon → mounted test. Each step retains the same identifier or a recorded child identifier. If a blank is recut or reworked, keep the earlier state and note the change rather than silently reusing its original dimensions. A measured final dimension and a drawing’s nominal dimension serve different purposes.

In the illustrative final C1 record, overall length is 60.0 mm and each grip engages 10.0 mm, leaving a free grip-to-grip span of 40.0 mm. The two specimen gauge marks are only 20.0 mm apart, centred within that span. Gauge length here means that measured mark separation, not the overall length or grip separation. Width and thickness are measured at the three stated gauge positions after preparation. Retain those local readings; a single average would hide a narrow or thin region.

The local measurements below are reference dimensions at 23 °C before the heated test. They are not asserted to remain identical after heating or loading. If a later calculation needs dimensions at a different state, that state needs appropriate evidence or a justified model. This lesson preserves the record; it does not prescribe a universal specimen shape or decide which area an unrelated test method should use.

Match conditions and their stability

Temperature, humidity, alignment, gripping and loading rate can influence the result. State what was controlled, what was measured and what varied. A nominal machine setting is not proof that the specimen experienced precisely that condition.

For the following separate heating record, the heater setting stays at 80 °C. Three validated specimen-temperature channels observe the marked left, centre and right gauge positions. Their placement, attachment and response checks are part of the record. Contact sensors exchange heat with the object and surroundings, and they take time to respond. A calibrated sensor at the chamber wall still does not directly establish the specimen’s interior temperature.

To make the example decidable, invent this record check before inspecting the values: at a listed time all three readings must lie from 78.0 to 82.0 °C inclusive, their largest-minus-smallest spread must be at most 2.0 °C, and each must have changed by at most 0.5 °C from its reading exactly five minutes earlier. The channels and their suitability are supplied as established for this arithmetic. These chosen limits are not a standard method, an uncertainty budget or a claim that every material is equilibrated after a fixed wait.

At 10 min, the coldest reading is only 77.5 °C and the spread is 2.5 °C, so the check fails. At 15 min all readings lie in range and the spread is 1.0 °C, but the left position has warmed by 1.5 °C since 10 min, so that row still fails. At 20 min, the spread is 80.1 − 79.4 = 0.7 °C and the three five-minute changes are 0.4, 0.2 and 0.1 °C. This is the earliest listed row that passes every condition.

Passing this sampled check is narrower than proving an exactly uniform temperature throughout the coupon or identifying the first passing instant between rows. Unsensed depths and times remain unresolved. Record the actual loading time and continue the relevant observations: a condition reached before loading may change as grips conduct heat or deformation occurs. Likewise, room humidity is not itself a measurement of moisture content within a polymer specimen, and a crosshead-speed setting is not automatically a measured gauge-strain rate.

A fixture can change the deformation being observed

A tensile fixture should transmit the intended load through the coupon, but the specimen and grips form a mechanical system. The clamped ends restrain motion differently from the free gauge surfaces. An offset or tilted load path can add bending to extension. Grip slip changes how motion is transmitted. These effects are not cured merely by having an accurate force channel.

A small paired-surface illustration makes that distinction visible. This is a separate idealised strip model, not a reconstruction of C1 from its dimension table. At a given load, compare simultaneous longitudinal strains at matching gauge positions on the two opposite broad faces. In a uniform, symmetric strip at common temperature, with small deformation and an approximately linear strain variation through thickness, write ε_front = ε_mid + b and ε_back = ε_mid − b. Here ε_mid is the centre-plane axial strain and b is the signed bending contribution at the front face; b is a strain, not a length. Adding and subtracting gives ε_mid = (ε_front + ε_back)/2 and b = (ε_front − ε_back)/2.

In separate invented observations, record A has front/back strains 0.050% and 0.050%. Record B has 0.080% and 0.020% at the same force, nominal geometry and material direction. Both have ε_mid = 0.050%, but B has b = 0.030%, or 0.000300 as a dimensionless strain. Looking only at B’s front face would report a value 60% larger than A’s front face and could be mistaken for a changed uniform material response. The pair instead reveals an asymmetric deformation under the stated model.

The pair is a diagnostic, not a complete alignment certificate or an automatic correction of a failed test. It cannot exclude twisting, bending in another plane, nonlinear through-thickness strain, unmatched gauge positions or a surface-state/temperature difference. In an actual investigation, check those alternatives and the grip/load path. Do not simply average two inconsistent readings and declare the intended loading established.

Match what should be common and name the intended difference

Suppose the question is whether direction changes the response of the prepared S7 sheet. A useful comparison keeps the stock identity, relevant preparation, final geometry, specimen conditioning and test path matched as far as the question requires, while deliberately changing the load direction relative to rolling. A remaining location difference belongs in the interpretation. One pair from different locations does not establish a general direction effect across all sheets.

If the question instead concerns a surface treatment, changing that treatment is intentional. Hold the other relevant conditions common and compare the treated state explicitly. Preparation is not always a mistake to remove; it can be the feature being investigated. The record should make the intended difference distinguishable from accidental differences.

Consider an incomplete comparison: P is an S7 parallel strip with finished edges, loaded when the 20 min temperature row passes; Q is from S8, direction unknown, one face polished, with only an 80 °C heater setting and a different grip spacing recorded. If their responses differ, the data cannot assign the difference to direction, stock or polishing individually. The repair is to recover the missing evidence if it exists and make a matched follow-up, not to rewrite Q’s missing temperature or direction as if it had been measured.

A compact record therefore connects identity and location, material/load direction, preparation and storage history, actual dimensions and measurement positions, grips/restraints, actual condition observations and times, loading control and response method, and any departure or unusual event. Link those records to the individual response file. More repeated readings cannot reconstruct a lost origin or make different tested states equivalent. The calibration, reproducibility and sampling lessons address separate parts of that evidence chain.

Worked examples

Try the reasoning

1. Why record rolling direction?

   Show the answer and reasoning

   The material may be anisotropic.

   Different orientations can give different responses.

2. Can polishing affect a surface-sensitive result?

   Show the answer and reasoning

   Yes.

   It changes the tested near-surface condition.

3. A new 60.0 mm coupon is gripped for 8.00 mm at each end. Its centred 20.0 mm gauge is unchanged. Find the new free span and mark-to-inner-grip distances. Is its mounting identical to C1?

   Show the answer and reasoning

   The free span is 44.0 mm; each clearance is 12.0 mm. The mounting is different.

   60.0 − 2 × 8.00 = 44.0 mm and (44.0 − 20.0)/2 = 12.0 mm. The gauge length can match while the surrounding restraint and grip engagement change; retain both records rather than calling the entire geometry identical.

4. In a separate valid observation pair, 25 min readings are 79.3, 79.6, 79.8 °C and 30 min readings are 79.8, 80.0, 80.1 °C. Apply the same inclusive record check at 30 min. What remains unresolved?

   Show the answer and reasoning

   The 30 min row passes: spread 0.3 °C and changes 0.5, 0.4, 0.3 °C. Unobserved positions and times remain unresolved.

   All three final values lie in 78.0–82.0 °C. Equality at the 0.5 °C change limit is allowed by the stated rule. This passing sampled pair does not prove a whole-volume equilibrium or a service-temperature rating.

5. At a later 35 min reading the same heater setting remains 80 °C, but specimen readings are 79.9, 80.0, 82.3 °C. Does the earlier passing row make this new row acceptable?

   Show the answer and reasoning

   No. The right value exceeds 82.0 °C and the spread is 2.4 °C.

   Both the temperature-range and spatial-spread conditions fail, independently of the change check. A previous condition record does not establish the later state. Retain the deviation and investigate rather than replacing the measured 82.3 °C with the heater setting.

6. A new paired-surface record gives front 0.090% and back 0.030% under the same stipulated linear model. Find ε_mid and b. Has the centre-plane strain increased by the same factor as the front reading relative to A?

   Show the answer and reasoning

   ε_mid = 0.060% and b = 0.030%. The centre-plane ratio to A is 1.20, while the front ratio is 1.80.

   Average and half-difference separate the two terms: (0.090 + 0.030)/2 = 0.060 and (0.090 − 0.030)/2 = 0.030, in percent units. These are different observations. A material explanation for the 20% centre-plane difference still requires matched conditions and the model’s validity.

7. A file contains “extension 0.010 mm” and a photograph of 60 mm-long C1, but the tracked points are not identified. Can the gauge be reconstructed just by dividing by 60 mm?

   Show the answer and reasoning

   No. Overall length is not evidence of the measurement interval.

   If the extension belonged to the 20.0 mm gauge, the relative change would be 0.0500%; if it belonged to the 40.0 mm free span it would be 0.0250%. Neither interpretation is established by the photograph alone. Recover the actual tracked-point record or state that it is missing.

8. P is parallel to S7’s rolling direction with finished edges and measured specimen temperatures. Q is polished, direction unknown and described only as “80 °C”. Their response curves differ. Propose a follow-up that tests polishing rather than silently changing several factors.

   Show the answer and reasoning

   Use identified comparable stock and direction, common final geometry and mounting, and the same verified specimen-condition path; deliberately vary and record the polishing treatment.

   Preserve which face and how much material was removed, relevant preparation/storage times and local dimensions after treatment. Keep the original unmatched records as such. A matched follow-up can test the treatment association; neither the existing pair nor one new pair automatically establishes all batches or a unique microscopic cause.

What to carry forward

• Specimen preparation belongs in the evidence.
• Record actual conditions where they affect interpretation.
• Nominal drawings, measured reference dimensions, grip separation and specimen gauge length name different information.
• A chosen sampled condition check does not prove exact whole-specimen equilibrium.
• A paired-surface diagnostic can expose deformation that a single reading hides.
• Name the intended difference and retain the other relevant state and procedure records.