Author: Webster Publish Time: 2026-11-18 Origin: Site
"Grade 8.8" Doesn't Mean High Strength — and "Zinc Plated" Is Where Rework Hides: A Buyer's Guide to Racking Connection Components
Four things the short line on a racking quotation hides: what the grade number actually encodes, which coating standard you should be naming, who decides the salt-spray hours, and the hydrogen-embrittlement trap created by combining hot-dip galvanising with high-strength grades.
This guide maps Chinese national standards (GB) to their ISO bases, because nearly every coating and fastener standard cited below is an adoption of an ISO document. Where the two documents are equivalent, you can specify either.
The governing document is GB/T 3098.1-2010, Fasteners — Mechanical properties of bolts, screws and studs, in force since 10 January 2011, replacing the 2000 edition and adopting ISO 898-1:2009 with modifications. Scope: carbon or alloy steel fasteners, coarse thread M1.6–M39 and fine thread M8×1–M39×3, tested at 10–35°C.
The designation is two digits separated by a dot:
• Digit left of the dot × 100 = nominal tensile strength in MPa
• Digit right of the dot ÷ 10 = yield ratio (yield strength / tensile strength)
• Multiplying the two gives nominal yield strength
Property class | Nominal tensile | Nominal yield | Typical racking use |
4.8 | 400 MPa | 320 MPa | Shelf-panel bolts, trims and guards — low-loaded positions |
6.8 | 600 MPa | 480 MPa | Light-duty shelving connections |
8.8 | 800 MPa | 640 MPa | Frame foot plates, beam-to-column connections, guard rail fixing |
10.9 | 1000 MPa | 900 MPa | High-bay frame bases, seismic connection nodes |
Three traps:
A higher number is not better quality. The 8.8 / 10.9 / 12.9 stamped on a bolt head denotes a strength class, not a quality ranking. What a racking joint actually needs is stable preload and batch-to-batch consistency; going up a grade can introduce new problems (see section 5).
Same class, different performance. Bolts of the same class produced by different manufacturers to GB/T 3098.1 can differ substantially in practice, particularly in fatigue performance; up-forged heads and rolled threads outperform cut threads on fatigue. This is the entire variance that the three characters "8.8" conceal.
De-rated product must be marked "08.8". Where material properties match class 8.8 but the product's actual load-carrying capacity is lower, the standard requires the property class to be designated "08.8". If you see 08.8 on packaging, it is not a printing error — it is a de-rated fastener.
Testing tensile strength alone verifies roughly a third of the picture.
Proof load. Proof stress is proof load divided by the stress area; the test requires the fastener to hold the specified load for 15 seconds without fracture or permanent deformation. This is the cheapest and least dispensable ex-works test in the book.
Core hardness versus surface hardness. Hardness reveals whether heat treatment was done properly — over-hard means brittle, under-hard means insufficient treatment. Methods run to GB/T 230.1-2018 (Rockwell), GB/T 231.1-2018 (Brinell) or GB/T 4340.1-2024 (Vickers). Where you measure matters: core hardness is taken on the thread-end cross-section or a rod section and represents bulk strength, while surface hardness reflects only the outer layer. A wide gap between the two points to surface decarburisation or carburisation — the starting point of early thread fatigue cracking.
Decarburisation depth. GB/T 3098.1-2010 sets an upper limit on decarburisation depth; a published failure analysis used a measured depth exceeding the standard's stated maximum (15 μm) as its evidence of an out-of-control heat-treatment step. Most racking bolt failures begin at the thread root, and decarburisation is the first hazard sitting there.
Hardenability also governs hardness distribution: poor hardenability lowers core strength, which is why two mills working from your identical drawing can diverge in test results.
There are at least three distinct GB coating standards, covering three different processes, service lives and price points.
Standard | Title | Process | Status |
GB/T 5267.1-2023 | Fasteners — Electroplated coatings | Electroplated zinc, zinc alloys (Zn-Ni, Zn-Fe), cadmium | Issued 2023-05-23, in force 2023-12-01; adopts ISO 4042:2022 with modifications; replaces the 2002 edition |
GB/T 5267.2-2021 | Fasteners — Non-electrolytically applied zinc flake coatings | Zinc flake (Dacromet-type) | Issued 2021-12-31, in force 2022-07-01; replaces 2017 edition; adopts ISO 10683:2018 with modifications |
GB/T 5267.3-2008 | Fasteners — Hot dip galvanized coatings | Hot-dip galvanising | In force 2009-02-01; identical adoption of ISO 10684:2004 |
Sources for the designations and dates are the public standards records. ISO 10684:2004 remains current, last reviewed and confirmed in 2021.
Why the standard number belongs on the drawing. Where a drawing says only "zinc plated, passivated", one supplier delivers 5 μm thin-passivate and another delivers 8 μm thick-passivate plus sealant — a cost gap of more than 30% — while the two quotations look identical. The difference lives in coating thickness, passivation type and whether a sealant is present.
Two dimensional constraints on electroplated coatings are worth separating out. First, each nominal coating thickness corresponds to a minimum local thickness and a batch-average range — the published thickness table pairs a 3 μm nominal with a 3–5 μm batch average, and so on. Second, to limit interference risk in the thread, coating thickness must not exceed one quarter of the basic thread tolerance. Hot-dip galvanising deposits a much thicker coating, which is why GB/T 5267.3 specifies separately how nuts are tapped oversize and bolts supplied with undersized threads.
Version caveat worth flagging: the thickness-table values and the one-quarter rule above are reproduced from publicly available extracts of the 2002 edition of GB/T 5267.1. The controlling document is now GB/T 5267.1-2023. Confirm the figures against the current text before writing them into a contract or drawing.
This is the most widely misunderstood item in the category.
The test method itself is standardised. GB/T 10125-2012 (neutral salt spray, NSS) specifies 5% sodium chloride solution, pH 6.5–7.2, 35°C ± 2°C, continuous spray, at a fallout rate of 1.0–2.0 mL per 80 cm² per hour, and specimens held at 15°–25° from vertical; rating follows methods such as ISO 10289. The international counterparts are ISO 9227 and ASTM B117.
But "how many hours" is not set by any standard — it is agreed between buyer and supplier. Published testing-laboratory guidance gives the customary working ranges: electroplated zinc (5–15 μm) typically 72–120 hours, hot-dip galvanised (40–80 μm) typically 240–480 hours. A different laboratory gives red-rust-free endurance of 72 hours or more for blue-white passivated electroplated zinc, and 200 hours or more for hot-dip galvanised. The two sets disagree; both are empirical ranges, not standard clauses.
Write the pass/fail rule into the contract:
• Red rust (substrate corrosion) = fail. This is a one-strike criterion.
• White rust (zinc corrosion product) = commonly accepted up to 5% of area, though some buyers require none.
Practical conclusion: put a number and a criterion on your drawing — "X hours with no red rust, white rust ≤ 5% area" — rather than "to national standard". The standard will not choose the duration for you.
This combination is unusually relevant to racking, because outdoor and cold-chain racks routinely require hot-dip galvanising while frame bases commonly use class 10.9.
The mechanism: during pickling and cleaning before hot-dip galvanising, hydrogen can be absorbed into the steel and cannot be fully expelled. The higher the strength class, the greater the sensitivity to hydrogen-induced delayed fracture. GB/T 5267.1-2023 now places internal hydrogen embrittlement requirements in the body of the standard (clause 4.4) and adds recommendations for reducing the risk in Annex B.
The zinc flake standard takes a different route: GB/T 5267.2-2021 contains a mandatory prohibition on pickling for high-hardness and high-class fasteners.
Source discrepancy — reported, not resolved: published descriptions of that pickling ban give the hardness threshold as either "above 390 HV" or "above 385 HV" for class 12.9 and above. We are not selecting one for you. Verify against the authoritative text before signing.
Three practical rules:
1. Specify 8.8 unless the calculation demands 10.9. A class 10.9 hot-dip galvanized frame base is among the highest-risk hydrogen-embrittlement combinations in a racking system.
2. Require the de-embrittlement process record. Bake temperature and hold time after galvanising belong in the technical agreement, not in a verbal understanding.
3. For critical loaded parts, require delayed-fracture or proof-load re-testing. GB/T 5267.3 permits stress relief to be specified before pickling and hot-dip galvanising where the application justifies it.
Buyers are often told "make it to the rack standard". The actual mapping is narrower than that phrasing suggests.
Standard | Title | Relevance to components |
GB/T 28576-2012 | Calculation of industrial rack design | Covers assembled industrial racks: basic structure, calculation model, load cases and combinations, and strength, stiffness and stability verification. Reviewed in 2024 with the conclusion to remain valid |
GB/T 27924-2011 | Industrial racks — Dimensions and rated loads | Sets the dimensions and load classes your components must match |
GB/T 33454-2016 | Operation specification of storage rack | End-use requirements, which drive wear-part replacement frequency |
GB/T 28576-2012 is administered by the National Technical Committee of Logistics Standardization (TC269) under the Standardization Administration of China; lead drafters include the Logistics Research Institute of University of Science and Technology Beijing, Nanjing INF Storage Equipment Engineering, Shanghai Dinghu Industrial Equipment and Jiangsu Sixin Logistics Equipment.
The operative point: rack standards define loads and verification methods. They do not specify your connection component's property class, coating thickness or salt-spray duration. Those three must be set by the rack integrator on the component drawing, or confirmed back by the component supplier at quotation stage. Whoever supplies the complete parameter set first holds the advantage in the negotiation.
Two small parts, for completeness: retaining pins (split pins) are supplied to GB/T 91-2000, covering nominal sizes 0.6–20 mm, in carbon steel (Q215, Q235), stainless (Cr17Ni7, 0Cr18Ni9Ti) or copper alloy (H63); installation requires the opened legs to exceed 180° (or 60° where interference dictates), leg offset no greater than the nominal size, and explicitly forbids substituting wire. Rack safety pins are a separate standard part, typically 3/8 mm nominal diameter, with the pin body in stainless or ordinary carbon steel.
4. Property class to GB/T 3098.1-2010 (ISO 898-1), class stated explicitly; de-rated product marked per the standard (e.g. 08.8).
5. Per lot, supply: proof-load test record, measured core and surface hardness, decarburisation measurement — all traceable to the heat/lot number on the mill certificate.
6. Coating stated as four elements — standard number, nominal thickness, passivation type, sealant yes/no. Two words like "zinc plated" are not acceptable.
7. Salt spray stated as method plus duration plus criterion: to GB/T 10125-2012 (ISO 9227 / ASTM B117 equivalent), X hours with no red rust and white rust ≤ 5% area.
8. For class 10.9 and above that are hot-dip galvanised: attach the de-embrittlement record (bake temperature, hold time) and re-test as agreed.
9. After coating, verify go/no-go gauge on the thread to confirm coating thickness stayed within one quarter of the basic thread tolerance.
Standard titles, numbers, issue and implementation dates, and the degree of ISO adoption are taken from the public national standards information service and the national standards platform. Industry ranges (salt-spray hours in particular) come from third-party testing-laboratory publications and are empirical, not mandatory clauses — this distinction is flagged at each occurrence. Where sources disagree (the 385 HV versus 390 HV pickling threshold; the two salt-spray ranges), both readings are shown rather than one being chosen. The coating-thickness figures in section 3 come from public extracts of the superseded 2002 edition and must be confirmed against GB/T 5267.1-2023 before contractual use. Confirm every clause against the authoritative standard text before signing.