Importing Industrial Fasteners From China: Torque, Coating and the Grade Nobody Checked

I have a drawer in my office with about forty bolts in it. All of them came back from suppliers that claimed to be 8.8 grade. Three of them are, in the functional sense. The drawer is the reason I no longer accept a property class stencil as evidence of anything.

Fasteners look like the simplest thing you can buy from China. Threaded rod, hex bolts, nuts, washers. There is no firmware, no battery, no certification maze. That simplicity is exactly what makes the category dangerous, because buyers stop asking questions earlier than they would for a power supply.

Property class is a mechanical claim, not a label

A 8.8 bolt means a nominal tensile strength of 800 MPa and a yield ratio of 0.8, so roughly 640 MPa yield. On a M10 coarse thread that is around 46 kN of proof load. Those numbers come from the steel and the heat treatment, and both can be quietly downgraded while the head still gets stamped 8.8. I have seen bolts marked 8.8 that sheared at somewhere between 60 and 70 percent of the proper proof load, which is what happens when the plant buys low-carbon wire rod and skips a step in the quench and temper line to save a day.

What actually catches this is a tensile test on a sample pulled from the production lot, done by a lab that is not the factory's own. It costs very little, maybe 30 to 80 USD per lot depending on the lab and the number of samples, and it is the single most useful thing you can buy in this category. Ask for the raw load-deflection curve, not a certificate that says "pass". The curve shows you whether the bolt yielded where it should have or snapped brittle, and brittle failure is the signature of a bad heat treat.

For anything that carries load rather than just holds a cover on, I also want a hardness map across the cross-section. Surface hardness can be fine while the core is soft, and the core is what fails under shock.

Hydrogen embrittlement is the delay bomb

High-strength fasteners at class 10.9 and above, particularly when they are electroplated with zinc, can absorb hydrogen during pickling and plating. The part passes every test in the factory, ships, gets installed, and cracks three weeks to six months later under sustained load. It is a delayed failure, so your incoming inspection will never see it.

The controls are boring and specific. Ask whether the plating line does a post-plate bake within the window your standard calls for, and for high-strength zinc-plated parts that usually means within a few hours of plating at around 200 to 230 °C for several hours. ASTM B850 is the reference document most suppliers will recognise, and a plant that has never heard of it is a plant that also has never had a customer catch them. Mechanically applied zinc coatings or zinc flake systems sidestep the problem entirely, which is why automotive work moved that direction years ago.

Coating changes your torque numbers

This is the one that quietly ruins assemblies. The torque you need to reach a given preload depends on the friction coefficient, and the friction coefficient depends entirely on the finish. A plain oiled black oxide bolt might sit around k=0.16 to 0.20. Zinc plating with a wax topcoat can land near 0.14. A zinc flake coating with a lubricant built in can go lower still.

If your engineering drawing specifies a torque value that was developed on an uncoated sample and you switch the supplier to a plated part, you can over-stretch the bolt or under-clamp the joint without anything visibly going wrong. I have watched a customer chase a leaking gasket for two months that turned out to be a coating change moving the clamp load by 15 percent.

Practical fix: whenever the finish changes, ask for a torque-tension test (the ISO 16047 method is the common one) and set the torque from that data rather than from a table. Good fastener plants will have a torque-tension rig; the ones that don't will quote you the number from a handbook and hope.

Thread and dimensional checks that actually catch problems

Thread pitch diameter is where cheap production hides. Go/no-go gauges are standard and any decent plant owns them, but the question is whether they are used on every lot or pulled out when a customer visits. Ask for the gauge records with the lot number on them. If the records are all dated the same afternoon, you have your answer.

For nuts, check the width across flats and the thickness. A nut that is 0.5 mm thinner than spec will strip the bolt threads under load long before the bolt itself fails, and the failure will be blamed on the bolt. Metric classes 8 and 10 for nuts are the counterparts to 8.8 and 10.9 bolts and the pairing matters; a 10.9 bolt with a class 8 nut is a mismatch that shows up in vibration testing and not much else.

What separates a plant from a trader

Fastener trading is extremely common out of Ningbo, Wenzhou and Yongnian, and there is nothing wrong with a good trader. The problem is when you think you are buying from the plant and you are three layers up. Three questions usually resolve it in one conversation:

• Which cold heading machines do you run, and what is the maximum diameter? A plant will name models and tonnage without hesitation. A trader will talk about their partner factory.
• Where is your heat treatment done, in house or subcontracted? Subcontracted is normal at smaller plants, but they should know the name of the subcontractor and be able to show you the furnace load records.
• Can I buy a lot with my own lab's witness at the tensile test? Traders hate this question because it exposes which plant actually made the parts.

None of these are trick questions. A straightforward supplier answers them in under a minute. The evasive answers are the data.

Reasonable first order

I would not place a container of fasteners on the strength of a sample bag. Start with the highest-load application you have, order maybe 2,000 to 5,000 pieces, pay for independent tensile and hardness testing, and torque-test ten assemblies to failure so you know where the real limit sits. That testing typically adds a few hundred dollars to a first order and it has saved me far more than that in the field.

One last thing worth stating plainly. If a quoted price is 30 percent below everyone else for the same grade and finish, you are not being offered a bargain, you are being offered a different product with the same stencil on the head. The drawer in my office exists because of exactly that.