A steel mill certificate says the material was tested to ASTM A370. A laboratory opens ASTM E8 and machines a dog bone. Both parties think they are talking about the same test, and most of the time the result comes out close enough that nobody looks further. The gap shows up on the day a customer rejects a heat, and the argument turns on which specimen was cut and where it came from.
ASTM A370 is the mechanical testing framework used for steel products alongside the applicable material or product specification. A370 defines test procedures, terminology and standard specimen configurations, while the pertinent product or general requirement specification normally controls sampling frequency, specimen location and orientation, material condition and reporting. That distinction matters before machining begins: the preparation team needs both the test method and the product specific sampling requirements. This guide covers what that means at the machine, from the blank to the finished specimen.
What A370 Adds on Top of a Tension Test Method
A tension test method explains how a tensile test is performed and how its results are determined. ASTM A370 places mechanical testing in the context of steel products and provides standardized procedures for tension, bend, hardness and impact testing. The applicable product specification then determines which tests are required, how often they are performed and how the test material is sampled.
For specimen preparation those documents have to be read together. A specimen drawing can define the finished geometry, but the preparation team still needs the product specific requirements that identify the permitted specimen type, sampling location and orientation. This is also why a fabricator quoting A370 often needs flat tensile blanks, round specimens and Charpy blanks from the same piece of steel.
Practically, that means the preparation shop cannot work from the tensile drawing alone. The drawing tells you the shape. A370 tells you which shape you are allowed to use for that product, and where the material has to come from.
Sampling Decides More Than Machining Does
The most expensive mistakes in specimen preparation are not dimensional. They happen when a correctly machined specimen was cut from the wrong place.
Rolled products behave differently along the rolling direction and across it, so orientation is part of the specification rather than a detail. Thick sections are not uniform through the thickness, which is why the sampling position, mid thickness or quarter thickness, has to be recorded. Heat treated products carry a gradient from the surface inwards. In each case the specimen is a claim about a region of the product, and the paperwork has to say which region.
This is worth building into the preparation routine itself. Mark the blank with the orientation and the position before it leaves the cutting station, because once it is a rectangle on the bench, nobody can tell which way it came out of the plate.
Test frequency belongs in the same conversation. A product standard decides how many tensile and impact specimens a heat or a lot owes, and that number drives the preparation schedule far more than any single drawing does. A laboratory that machines one specimen at a time because the request arrived one at a time ends up doing the same setup five times for the same heat.
Full Section, Strip and Round, and When Each Applies
A370 recognises that steel products do not all arrive as convenient plates, and the specimen options follow from that.
Full section specimens keep the product as it is, which is common for smaller bars and for tube where the whole cross section can be pulled. Nothing is machined away, so the result reflects the product rather than a machined core, but the grips have to accept the real shape and the machine has to have the force to break it.
Strip specimens are cut from flat rolled products, sheet and plate, and machined to a rectangular gauge. This is the workhorse for structural steel work and the format most preparation machines are built around.
Round specimens are turned from thicker product where a full section pull would need more force than the frame has, or where the standard calls for a machined round. The gauge diameter and the shoulder or thread form follow the standard and the grips that will hold it.
Tube brings its own decisions. Depending on the product and the specification, the test may use the full tube section, a longitudinal strip taken from the wall, or a machined round taken from the wall thickness. Each of those needs different preparation and different fixturing, and the choice belongs in the method rather than on the shop floor.
Bend and Impact Blanks Come From the Same Coupon
A370 is not only a tension document, and a laboratory that treats it as one ends up cutting the same plate three times.
Bend test specimens are usually rectangular blanks with the edges dressed, and the detail that decides the result is the edge condition rather than the width. A sheared edge with a work hardened lip starts a crack that has nothing to do with the steel, so the edges are machined or ground and the corners are lightly rounded before the specimen goes into the bend fixture. Thickness matters too, because the bend radius in the specification is written against the specimen thickness, and a blank machined thinner than intended is being bent more severely than the standard asked for.
Impact blanks are a machining job in their own right. Full size Charpy specimens are 10 by 10 mm in section, and when the product is too thin to yield that, the standard allows reduced thickness specimens with the absorbed energy reported against the actual size rather than converted. The notch is cut last, after the bar is finished to size, because a notch machined into an unfinished blank moves as the surrounding material comes off.
Our TensileMill CNC Classic Upgrade was built as a hybrid for exactly this reason: the same machine that produces flat tensile specimens also prepares impact blanks, so a coupon that owes tension, bend and impact results does not have to travel between three setups. The MICRO does the same work at the small end, on strip and thin sheet.
Where A370 and E8 Sit Beside Each Other
For flat and round machined specimens the dimensional detail in A370 lines up with the familiar geometry from ASTM E8 and E8M, so a laboratory preparing to one is rarely far from the other. The differences that matter are elsewhere.
E8 is written for metallic materials in general and concentrates on how the tension test is performed. A370 is written for steel products and concentrates on what has to be tested, how often, and from where. A specification that cites A370 is telling you about sampling and product forms. A specification that cites E8 is telling you about the test itself. Many jobs quote both, and a preparation programme that stores geometry by standard rather than by customer makes that painless.
The practical rule for a preparation cell is simple. Keep the geometries as saved profiles, name them by the standard and the product form, and let the operator pick rather than type. That removes the most common preparation error, which is a gauge length or a width entered from memory.
Reinforcing Bar and Other Full Size Product Tests
Reinforcing bar is the clearest case where machining a neat specimen would be the wrong answer. The product is tested as it is delivered, deformations included, because the deformations are part of what the customer is buying. That shifts the whole problem from the preparation shop to the grips and the frame.
Three things decide whether a full size test will work. The frame needs the force, which for larger bar sizes leaves the benchtop range entirely. The grips need jaw faces that hold a ribbed surface without shearing the ribs off, since a bar that slips reports a low yield and a long elongation. And the gauge length has to be marked on the bar itself, because there is no machined reduced section to define it.
The same logic applies to other full size product tests. Wire, strand, small diameter tube and finished components are often specified as full section pulls, and in each case the preparation work is cutting to length, cleaning the ends and marking, rather than machining a profile. Laboratories that plan for this buy the gripping first and the frame second, which is the opposite of the usual order and the correct one.
Preparing the Blank Without Damaging the Material
Everything that touches the specimen before the mill leaves something behind. Shearing work hardens the edge. Flame and plasma cutting leave a heat affected zone. Laser cutting leaves a narrower one. Waterjet leaves the least thermal damage but still needs stock removed.
The rule that keeps results honest is to leave enough material for machining to remove all of that, then take it off with the cutter rather than hoping it does not matter. On thin sheet the affected band is small; on thick plate cut with flame it is not, and a specimen machined too close to the cut edge reports properties that belong to the cutting process rather than to the steel.
Heat during machining deserves the same respect. Steel does not need to glow to be affected, and a specimen that has been pushed hard with a dull cutter carries residual stress into the test.
Machining Flat Specimens for Steel Products
Flat specimens are where most A370 work lands, and a purpose built machine earns its place because the geometry repeats thousands of times a year.
Our TensileMill CNC Classic Upgrade handles specimens up to 14 inches (350 mm) long and 2 inches (50 mm) wide, drives a 3.2 kW (3 HP) servo spindle, and works on material up to 60 HRC. Its published figures are a positioning accuracy of plus or minus 0.13 mm and a repeatability of plus or minus 0.02 mm, which covers the specimen geometries in common ASTM, ISO, DIN and JIS practice. With the triple clamping fixture it prepares up to three stacked sets of one inch thick material in a single setup, which is the difference between a shift of machining and an hour of it.
Throughput is the argument that convinces management, and it is worth stating in the units they use. A coupon that owes two tensile specimens, three impact blanks and a bend blank is a full morning of manual milling with setup between each shape. On a dedicated machine with saved profiles it is one setup and a run, and the operator spends the rest of the morning on work that needs judgment. That difference repeats every time a heat is released. The useful payback calculation is therefore built on setup time avoided, operator hours recovered, throughput gained and the cost of remaking specimens, rather than on machine cycle time alone.
Where floor space is short, the TensileMill CNC MINI runs the same TensileSoft workflow with saved specimen profiles and a 2.2 kW water cooled spindle. For higher volumes and heavier stock, the TensileMill CNC XL offers 400 by 230 by 400 mm travels, an 800 by 240 mm table, 150 kg load capacity and a 10,000 rpm spindle. Thin sheet and small product forms are the territory of the TensileMill CNC MICRO, which also prepares impact blanks from strip.
To 60 HRC, 8 specimens per run
Table 800 x 240 mm, 10,000 rpm
Soft, hard and Inconel stock
Metals, plastics and composites
Thin Product and Reduced Specimens
Sheet steel below a couple of millimetres brings its own set of decisions. The specimen stays flat, the thickness is whatever the product is, and the width and gauge length come from the standard rather than from convenience.
Two practical problems dominate. The first is holding the blank while it is machined, because a thin strip clamped over a large span lifts and chatters, and chatter marks along the reduced section are exactly the notches a tensile specimen should not have. Stacking several blanks and machining them together solves it, which is why stack capacity appears in our machine specifications at all: the Classic Upgrade takes up to three stacked sets in the triple clamping fixture, and the MICRO can produce either one thick specimen or several thin ones in a single run.
The second is buckling in the test, not in the machine. A thin specimen that is slightly wavy after machining will straighten under load before it starts stretching, and that shows up as a soft first part of the curve and a modulus figure that nobody believes. Flatness is a preparation requirement even though no standard prints a number for it on the machining drawing.
Turning Round Specimens From Bar and Heavy Sections
Round specimens come out of a lathe, and the same argument about repeatability applies. A general purpose lathe can make a beautiful specimen when a skilled machinist has time. A specimen preparation lathe makes the same specimen on Friday afternoon in the hands of a technician.
The TensileTurn CNC Classic Upgrade takes stock up to 1.5 inches thick and material up to 55 HRC, holds up to three stacked samples, and runs from the TensileSoft interface on a 15 inch controller. The TensileTurn CNC XL steps up for round, square and irregular stock in tougher materials, and the Industrial Upgrade is built to run independently of the general machine shop so that specimen work stops competing with production parts.
One detail that costs laboratories time is the end form. Threaded ends and shouldered ends both appear in steel work, and the choice has to match the grips waiting on the testing machine rather than the preference of whoever wrote the program.
Up to three stacked samples
Consistent geometry batch after batch
Flat and round specimens
Surface Finish, Radii and the Details That Move Results
Two features decide whether a specimen breaks where it should. The first is the transition radius between the grip section and the reduced section, which has to be smooth and continuous, because a step or a tool mark there concentrates stress and pulls the failure out of the gauge length. The second is the finish along the reduced section, which should run along the specimen rather than across it, since circumferential marks act as small notches.
Both matter more as the material gets stronger and as ductility drops. On mild steel a slightly rough gauge is forgiving. On a quenched and tempered grade it is not, and elongation figures start scattering for no reason anyone can see in the certificate.
For work where the finish has to be controlled rather than hoped for, the automatic longitudinal polisher takes the machining marks out in the axial direction and lowers the residual stress left by turning and milling.
Measuring the Specimen Before It Is Tested
The dimensions used in the calculation are the measured ones, not the nominal ones from the program, and this is where a good preparation cell quietly protects a laboratory.
Measure width, thickness or diameter exactly as the governing method requires, at more than one position in the reduced section, and use the original cross sectional area that the procedure specifies. Nominal dimensions from the machining program are not a substitute for the measurements the test method calls for. A specimen that tapers along the gauge is not automatically scrap, but the taper has to be inside the tolerance the standard allows, and the person testing it has to know. Round specimens are measured for diameter in two directions at the same station, since a slightly oval turned specimen reports a strength that depends on which way the caliper was held.
The habit worth building is measuring at the machine rather than at the testing frame. A specimen that is out of tolerance can be recut in minutes while the setup still exists, and it becomes an expensive nuisance once the batch has moved on and the operator has changed the program.
Building an A370 Routine That Survives an Audit
Auditors rarely ask whether the mill was set up correctly. They ask whether the laboratory can show what it did.
- Which specimen geometry was used, named by standard and product form rather than by nickname.
- Where the blank came from, including orientation and through thickness position.
- What removed the blank from the product, and how much stock the machining took off afterwards.
- Which program ran, and whether the saved profile has changed since that job.
- How the specimen was measured before testing, and by whom.
None of this is difficult when the preparation cell is set up for it. All of it is painful to reconstruct months later from memory.
Traceability travels with the blank as well. The heat or lot number, the coupon it came from and the specimen identity have to survive machining, and a marking that disappears with the first cut is no marking at all. Most laboratories settle on stamping or engraving in the grip section, away from the reduced area where a mark would act as a notch.
- Product form and specimen type, full section, strip or machined round.
- Orientation and position in the product, including through thickness location.
- How the blank was removed and how much stock the machining took off afterwards.
- The program or profile used, named by standard rather than by customer.
- Measured dimensions from the reduced section, not the nominal ones.
Preparation Errors That Show Up as Material Problems
When a tensile result looks wrong, specimen preparation and gripping are worth checking before the deviation is attributed to the material. Preparation errors are common contributors to off gauge breaks, slippage, scatter and inconsistent elongation, but no single result pattern is diagnostic on its own.
- Break outside the gauge length, usually a rough or stepped transition radius, a tool mark across the reduced section, or grip jaws biting into the shoulder.
- Low strength with high elongation, which is the signature of a specimen slipping in the grips rather than a soft heat.
- Scatter between specimens from one coupon, which can be associated with inconsistent width or thickness, or at specimens taken from different positions without being recorded as such.
- High strength on thin sheet, often a work hardened sheared edge that was never machined away.
- Elongation that varies with the operator, which can result from gauge marks being applied differently rather than the material behaving differently.
None of these are exotic failures. They are the reason a purpose built preparation cell pays for itself: not because a machinist cannot make a good specimen, but because the same specimen has to come out identical when a different person makes it on a different day.
Does ASTM A370 Define Tensile Test Specimen Dimensions?
ASTM A370 includes standard specimen configurations and dimensional requirements used for mechanical testing of steel products. The specimen permitted for a particular product can also depend on the applicable material or general requirement specification, including product form, thickness, sampling location and orientation. Use A370 together with the specification that governs the steel product being tested.
Deciding What Your Laboratory Actually Needs
Steel product work is rarely one specimen type. A plate mill needs flat strip specimens and bend blanks. A bar producer needs full section pulls on small sizes and machined rounds on large ones. A fabricator following A370 for structural steel usually needs flat specimens plus Charpy blanks from the same coupon.
Send us the product forms you test, the thickness range, the standards your customers cite, and how many specimens leave the shop in a typical week. We will come back with the machines and fixtures that cover that mix, and we will say plainly when one machine is enough, because a second machine that stands idle helps nobody.
The full range sits on the TensileMill CNC site, and if you also run the tests in house, the electromechanical testing systems are configured for the same standards.
TensileTurn CNC – Industrial Upgrade – Round Tensile Sample Preparation MachineThe TensileTurn CNC - Industrial Upgrade Model is a conveniently sized automatic CNC lathe for industrial use designed to prepare...View machine →
TensileTurn CNC XL – Round Tensile Specimen Preparation MachineThis robust lathe machining center for round tensile sample preparation is the ultimate solution for your manufacturing needs.View machine →
TensileTurn CNC – Classic Upgrade – Round Tensile Specimen Preparation MachineThe TensileTurn CNC – Round Specimen Preparation System has evolved into the newly upgraded Classic Upgrade Model , a robust and...View machine →
TensileMill CNC MICRO – Precision Flat Tensile & Impact Micro MachineThe TensileMill CNC MICRO is your compact, highly efficient solution for precise tensile specimen preparation.View machine →
TensileMill CNC XL – Flat Specimen PreparationTensileMill CNC XL - Flat Specimen Preparation is a powerful and dynamic addition to any manufacturing floor.View machine →
TensileMill CNC – Classic UpgradeExperience unparalleled ease of use with the TensileMill CNC - Classic Upgrade, the world's first flat tensile and impact test...View machine →
TensileMill CNC MINI – Compact Flat Tensile Specimen Preparation MachineCompact, high-power CNC machine for flat tensile and impact specimen preparation, smaller footprint, larger table,...View machine →
Automatic Longitudinal Polisher SystemThe Longitudinal Polisher TensilePolish from Material Testing Center is designed to lower grinding stress and surface residual...View machine →
Electro Mechanical Universal Testing System 5 kN / 50 kN with Precision Load CellThe TM-EML Series C - Dual-Column Benchtop and Floor-Standing Universal Testing System (5 kN - 50 kN) is an advanced...View machine →
Turning Inserts and Cutters for TensileTurn Round MachinesCarbide turning inserts and cutters for TensileTurn CNC round specimen preparation lathes. We stock a range for different...View machine →