If you buy machined, forged, or heat-treated metal parts, you’ve probably written “Rockwell dureté: 58-62 HRC” on a drawing and moved on. That’s not wrong — it’s just incomplete. Hardness, tensile, and Charpy impact testing measure three completely different things, and a part that passes one can still fail the application. This article breaks down what each test actually proves, what it costs, how long it takes, and how to write a spec that doesn’t leave a gap for your supplier to walk through.
Rockwell Hardness Testing: Fast, Cheap, and Frequently Misused
Rockwell hardness testing refers to a method that applies a minor load followed by a major load to a metal surface and calculates hardness from the resulting difference in indentation depth, rather than from an optically measured indentation diameter [Struers]. That depth-based approach is why it’s fast — no microscope, no operator judgment call on where an indent edge sits — which makes it the default choice for production floors and incoming-material screening [EmcoTest].
The catch is scale selection. HRC is standard for hardened steel and tool steel; HRB fits annealed steel, copper alloys, and softer aluminum; and surface Rockwell scales like HR15N or HR30N exist specifically for thin stock or shallow case-hardened layers where a standard indenter would punch through into the softer core [Struers, Industrial Physics]. Three standard bodies cover the method — ASTM E18, ISO 6508-1, and JIS Z 2245 — and all three define scales, indenter geometry, test force, specimen requirements, and reporting, but none of them tell you which scale to use for your part. That’s on you.
On cost, if the supplier hands over a flat, prepped coupon, budget roughly US$25-75 per test point at a typical North American third-party lab, usually with a lab minimum charge attached [InfinitaLab, IMR Test]. Turnaround is same-day to 1-3 business days for straightforward jobs [IMR Test, InfinitaLab]. That’s the cheapest, fastest test of the three — which is exactly why buyers over-rely on it. A part can hit 60 HRC and still fail in service if its tensile ductility or low-temperature toughness is out of spec, because hardness measures resistance to localized indentation, not tensile strength or fracture energy [Struers, ZwickRoell].
Action : Never write “Hardness: 58-62” on a drawing. Write “58-62 HRC per ASTM E18,” name the test locations, and require a minimum of three readings — otherwise your supplier can legally hand you an ambiguous number that isn’t tied to any recognized scale.
Tensile Testing: The Test That Actually Proves Strength and Ductility
Tensile testing refers to pulling a standardized specimen at a controlled rate until it fractures, generating a stress-strain curve from which yield strength, ultimate tensile strength (UTS), elongation, and reduction of area are derived [ZwickRoell, InfinitaLab]. Unlike hardness, this is destructive — the specimen is gone at the end — and it’s the only one of the three tests that directly answers “will this part hold load without permanently deforming or breaking?”
For materials without a sharp yield point, the industry standard is the 0.2% offset method, reported as Rp0.2 or “0.2% proof strength” [ZwickRoell, TensileMillCNC]. The governing method standards — ASTM E8/E8M, ISO 6892-1, JIS Z 2241 — tell you comment to pull the specimen; they don’t set a pass/fail bar. That comes from the material or product standard, most commonly ASTM A370 for steel, or the specific alloy/temper spec (think “4140 Q&T” or “6061-T6”) [ZwickRoell, ASTM]. I’ve seen buyers cite E8/E8M on a PO and assume that alone guarantees a minimum strength — it doesn’t. E8/E8M is a test procedure, not a strength requirement.
Here’s something I ran into directly while auditing supplier test reports for a batch of forged flanges: three different labs, using the same ASTM E8/E8M callout, returned wildly different-looking reports — one omitted sampling orientation entirely, one didn’t state whether the specimen came from the forging itself or a separate prolongation, and only one linked the specimen back to a specific heat number. All three were “technically compliant” with the standard, but only one report was actually useful for traceability if a part failed in the field later. That gap between “standard-compliant” and “useful for a real investigation” is the single biggest blind spot I’ve seen in tensile reporting, and it comes entirely from buyers not specifying report content up front.
Cost-wise, an already-prepared specimen runs roughly US$100-300 per piece; if the lab has to cut, turn, mill, or grind the specimen from raw stock, total cost commonly climbs to US$200-600+ per piece [InfinitaLab, Tensile.com]. Lead time is typically 2-5 business days for standard room-temperature testing, faster if specimens arrive pre-machined [SMS Labs, IMR Test].
Action : Require specimen traceability to heat/lot and heat-treatment batch in the PO, and for rolled plate, extrusions, forgings, or additive parts, explicitly call out sampling orientation (longitudinal, transverse, short-transverse) — orientation-driven property variation is real and standard callouts alone won’t catch it.
Charpy V-Notch Impact Testing: The One Buyers Skip Until Something Cracks
Charpy V-notch impact testing refers to swinging a pendulum into a notched specimen and calculating the absorbed fracture energy from the pendulum’s height loss, simulating how a notched component behaves under high-strain-rate loading — most relevant for assessing brittle fracture risk in steel, welds, and structural components at low temperature [ZwickRoell, World of Test]. Results are reported in joules or ft·lbf, sometimes alongside shear fracture percentage and lateral expansion, which help distinguish ductile from brittle failure modes [ZwickRoell, TestMetals].
The methods — ASTM E23, ISO 148-1, JIS Z 2242 — are consistent globally in structure, but the number that actually matters, minimum absorbed energy at a specified temperature, comes from the product or design specification, often paired with ASTM A370 for steel [TestMetals, standards.iteh.ai]. The parameter buyers forget most often is temperature. “Charpy impact ≥ 27 J” tells you almost nothing on its own; “27 J minimum at -20°C, three specimens, longitudinal orientation, per ASTM E23” tells you everything [World of Test, ZwickRoell]. Ferritic steels can lose toughness sharply as temperature drops, so a part that easily clears 27 J at room temperature can fail well below that threshold in a cold-climate or offshore application.
Standard full-size specimens are 10mm × 10mm × 55mm with a machined V-notch; where material thickness won’t allow a full-size specimen, sub-size specimens are used, but their results are not interchangeable with full-size data unless the governing spec explicitly defines a conversion or acceptance basis [World of Test, InfinitaLab]. This is also the test most sensitive to sampling direction — inclusion alignment, rolling texture, and weld heat-affected-zone microstructure all shift results by orientation, so “the material passed Charpy” is meaningless without stating where and in which direction the specimen came from [TestMetals, ZwickRoell].
Budget roughly US$150-450 for a standard set of three room-temperature specimens if you supply pre-machined coupons; add specimen extraction/machining, low-temperature conditioning (-20°C, -40°C, -46°C, -60°C), fracture appearance evaluation, witness testing, or ISO/IEC 17025 accredited reporting, and cost rises meaningfully [InfinitaLab, MVMTC, TestMetals]. Turnaround runs 3-7 business days for routine work, longer for low-temperature conditioning or specimen extraction from thick sections [IMR Test, InfinitaLab].
Action : For anything destined for cold service, pressure equipment, or structural welds, specify both the minimum test temperature and whether you require a single-specimen minimum or only a three-specimen average — labs will default to whichever is cheaper to report if you don’t say which one you actually need.
Cost and Lead Time: The Procurement Math Behind the Test Report
Testing cost and lead time refer not to how long the actual pull, indent, or pendulum swing takes — that part is minutes — but to how much specimen preparation, traceability, temperature control, and reporting the job requires [InfinitaLab, IMR Test]. This is the part of the process that catches procurement teams off guard, because RFQs often lump “testing” into one line item when the underlying cost drivers are completely different across the three methods.
| Cost driver | Rockwell | Tensile | Charpy V-notch |
|---|---|---|---|
| Base test | US$25-75/point | US$100-300/prepped specimen | US$150-450/set of 3 |
| Specimen prep impact | Low-moderate | Moderate-high | Moderate-high |
| Low/high-temp add-on | Rarely applicable | Can add significantly | Usually adds significantly |
| ISO/IEC 17025 reporting | Sometimes extra | Common extra charge | Common extra charge |
| [InfinitaLab, IMR Test, ZwickRoell] |
The industry-wide pattern is that routine tensile, hardness, and Charpy work quotes at 3-7 business days, with many accredited labs offering 24-48 hour expedited service at a premium [InfinitaLab, SMS Labs]. But “routine” is doing a lot of work in that sentence — non-standard geometries, temperature conditioning, and lab backlog routinely blow past those numbers. The opportunity here, if you’re dialing in a supply chain, is that separating “test fee” from “specimen machining fee” in your RFQ almost always exposes savings, because labs price specimen prep inconsistently and buyers rarely ask for the breakdown.
Action : Request testing and specimen-preparation costs as two separate line items in every RFQ, and get a written turnaround commitment before issuing a PO on anything with schedule risk — verbal “usually 3-5 days” estimates are not a delivery commitment.
Building a Test Spec That Actually Protects You
A defensible test spec refers to a procurement document that ties material grade, test method, sampling requirements, acceptance values, and reporting content together into one traceable package, rather than citing a test standard in isolation [Struers, ASTM]. The recurring failure mode across all three tests is the same: buyers cite the method standard (E18, E8/E8M, E23) and assume that alone sets the bar, when the method standard only tells the lab comment to test — the material or product standard sets the pass/fail line.
A spec that holds up should include, at minimum: material grade, condition, and governing product standard; the test method standard and version; sample location, quantity, orientation, and lot traceability; acceptance values with units (58-62 HRC, Rp0.2 ≥ 850 MPa, UTS ≥ 1,000 MPa, A5 ≥ 12%, CVN average ≥ 27 J at -20°C); required report content (individual values, average, test temperature, specimen dimensions, standard version, calibration status); and, for high-risk parts, a named third-party ISO/IEC 17025 accredited lab with defined witness and re-test rules [ASTM, ISO]. Hardness-to-tensile-strength conversion charts exist and get used informally in incoming inspection, but they shouldn’t substitute for an actual tensile pull on contractually critical material — chemistry, microstructure, heat-treat state, and anisotropy all break the assumption that equal hardness means equal tensile performance [Industrial Physics, Struers].
Action : Build a one-page test-spec template with these six fields as mandatory entries, and require your suppliers to fill in every field before a PO is released — a blank field is where ambiguity, and eventually a field failure, gets through.
FAQ
A: No. Hardness and tensile strength correlate loosely through empirical conversion charts, but chemistry, microstructure, and heat-treat state can break that correlation. Use hardness for fast screening and tensile testing for contractual strength verification, not as substitutes for each other.
A: Most often it’s a scale mismatch (HRC vs. HRB vs. surface Rockwell) or a test location too close to an edge, hole, or prior indentation, all of which distort readings. Always specify the exact scale and test locations in your PO.
A: Not automatically. Sub-size specimens are used when material thickness won’t allow full-size samples, but their absorbed-energy values aren’t directly interchangeable unless your governing spec defines an accepted conversion.
A: Test temperature. “Charpy ≥ 27 J” without a stated temperature is close to meaningless for ferritic steels, since toughness can drop sharply as temperature falls.
A: As a rough planning figure at typical North American third-party labs: US$25-75 per hardness point, US$100-600 per tensile specimen depending on prep, and US$150-450 per Charpy set of three — but these are budgetary estimates, not quotes, and should be confirmed via formal RFQ.
A: No. E8/E8M defines comment to run the tensile test; it sets no strength requirement. The minimum values come from the material or product standard (e.g., ASTM A370 or the specific alloy/temper spec), which must be cited separately.
