Two Numbers That Look the Same but Are Not
Open any MTC and you will see a chemistry block. What it represents depends on which analysis the mill performed, and the two are governed differently:
| Heat (cast/ladle) analysis | Product (check/verification) analysis | |
|---|---|---|
| Sample taken | During pouring of the heat — represents the whole ladle, typically 100–300 t of steel | From the finished or semi-finished product — pipe body, bar, forging |
| Represents | The chemistry the mill aimed for | The chemistry the product actually carries, including segregation effects |
| Governs acceptance | Yes — the spec limits apply as written | Yes, but with product analysis tolerances: slightly wider limits that absorb local heterogeneity and analytical scatter |
| Who does it | Always the producer | Usually on buyer request, during inspection, or in dispute |
Product analysis tolerances are allowance over the maximum or under the minimum — never a licence for the heat itself to be off-spec. A product reading outside the heat limits can still conform, if it stays inside the tolerance band; a heat reading outside limits is simply off-spec.
How the Tolerances Work (ASTM A751 Logic)
ASTM A751 defines the practice: the product analysis range is normally wider than the heat range, to allow for analytical reproducibility and the heterogeneity of the steel (micro-segregation is a fact of cast steel — elements like P and S segregate toward the centre of blooms and billets). The actual tolerance values live in the product specification or its general-requirements standard (for pipe, ASTM A530 and similar), element by element, and typically scale with the limit itself — a carbon max of 0.30% carries a larger product tolerance than a max of 0.06%.
- Product analysis may not vary both above and below the specified range for the same element across determinations.
- Tolerances cannot be used retroactively to accept a heat analysis that failed — different direction of use entirely.
- Chinese practice mirrors this: GB/T 222 specifies finished-product permissible deviations from heat-analysis limits for carbon and alloy steels.
Why This Matters at the Inspection Table
Suppose your inspector drills a delivered pipe and the lab reports C = 0.26% against a spec max of 0.25%. Instinct says reject; the correct first question is which analysis the spec applies. A product analysis has tolerance room above the heat maximum, so 0.26% may conform — while 0.29% would not. Walking through the logic in order:
- Confirm the sample was taken per the sampling practice (E1806 for steel products — location matters, decarburized surfaces skew C).
- Confirm the lab used referee-capable methods (E350 wet chemistry, E1019 for C/S/N; ICP acceptable for many elements).
- Apply the product-analysis tolerance from the product spec before judging pass/fail.
- If still failing, the spec’s dispute clause (re-sampling, referee lab) governs — not a unilateral rejection.
Certificates usually show heat analysis only. If your purchase order says ‘product analysis per lot’, say so explicitly and price it — it means drilling or machining product and adds a lab round. For carbon grades in ordinary service, heat analysis plus PMI screening is usually proportionate; reserve product analysis for alloy grades, C-Mo/Cr-Mo grades and disputes.
Practical Verification Menu
- PMI screening (XRF) — sorts grades, no carbon. Every heat at packing if risk warrants.
- Portable OES — adds carbon on the spot; small spark mark.
- Lab product analysis — the contractual answer; sample per E1806, test per A751 methods, judge with tolerances.
- Cross-checks — chemistry should cohere with hardness and tensile: a ’20#’ pipe reading like 45# steel will also machine and weld differently. Inconsistency between tests is itself evidence.
Frequently Asked Questions
Why is the product analysis range wider — isn’t that a loophole?
It is compensation for physics: cast steel is micro-segregated and surface sampling adds error. The heat analysis is the better representation of the whole ton; product analysis with tolerances is the fair test of the piece in front of you.
Do EN and GB standards work the same way?
Yes in principle: EN product specs allow specified variations for product analysis, and GB/T 222 sets finished-product permissible deviations vs the heat limits. Values differ from ASTM — always use the family’s own table.
Who pays for a disputed product analysis?
Per most general-requirements standards, the losing party pays: if the referee analysis shows conformity, the buyer pays; if not, the mill pays. Agree the lab up front.
Can I demand product analysis on every pipe?
Contractually yes, practically expensive. Sampling per heat per lot, plus PMI screening for identity, achieves most of the assurance at a fraction of the cost.
Does product analysis replace PMI?
No — PMI is fast screening over many pieces; product analysis is slow, accurate and definitive on few pieces. They answer different questions and work best chained together.
Standards & references: ASTM A751 · GB/T 222 · ASTM E1806 / E350
Need carbon steel pipe with full, verifiable inspection records? CREATEEL exports welded and seamless carbon steel pipe, fittings and coated pipe with complete MTCs and third-party inspection support. Talk to our team.
