VDI 2230 Bolted Joint Checks (Simplified) — Preload, Torque & Fatigue

Four M10 bolts, tightened to spec. And the bracket still slips.

This is VDI 2230 — the German guideline for high-duty bolted joints. Let's see how it works, field by field.

A bolted joint is two springs. Tightening stretches the bolt and squeezes the plates. That squeeze is the clamp force, and it holds the joint together.

When a load pulls on the joint, only a small share reaches the bolt. Most of it takes clamp away from the plates. VDI 2230 is the bookkeeping of that clamp.

Step one: the fastener. Thread size sets the stress area the bolt carries its load on — here an M10.

Property class sets the strength. Class 8.8 means at least 640 MPa before it yields — or choose Custom and enter your own.

Quantity is how many bolts share the load — four here, sharing it equally.

Friction μ is the friction in the thread. Enter the lowest you expect: most of the torque goes into friction, so a wrong μ means a wrong preload.

Does the head sit on a different surface — say a lubricated thread under a dry head? Tick the box and give the head its own friction, μ_K. It changes only the torque; the preload limit depends on the thread alone.

Rolled after heat treatment raises the fatigue limit — tick it only if the bolt maker certifies it. Standard bolts are rolled before.

Step two: the joint. Clamp length is everything squeezed between head and nut — 30 mm of steel.

Shank length, joint type and head type set how springy the bolt is — and a tapped joint has one bearing face instead of two.

A waisted bolt? Enter the waist diameter. A thin waist is a softer spring, which helps fatigue — but it becomes the weakest section, so the preload and the torque drop with it.

Surface roughness, interfaces and washers decide embedding: the surfaces bed in by a few micrometres, and that costs preload for good.

Every washer you count is drawn on the sketch — each one is another face that beds in.

Plate diameter and modulus set how stiff the plates are — aluminium is softer than steel. The bearing face and the hole come from the ISO tables unless you enter them.

From all of this the tool computes the load factor Φ = 0.157: the bolt feels about 16 % of any external load. Embedding costs 3.2 kN of preload per bolt.

Step three: tightening. This is where most joints are won or lost.

A torque wrench is cheap but imprecise: the real preload scatters. VDI expresses that as the tightening factor α_A — 1.6 here.

So the bolt is tightened to 29.6 kN at most — but you can only count on 29.6 divided by 1.6: 18.5 kN.

The tool gives you the torque to apply: 48 N·m. If your process is better controlled, enter your own scatter.

Step four: the loads. The axial load pulls the joint apart — 6 kN in total.

Does the load never drop to zero, or reverse? Tick the box and enter the minimum. Fatigue depends on the range — a reversing load cycles the bolt twice as hard.

The shear load slides the plates across each other — 7 kN. Only friction between the plates carries it.

That is the interface friction μ_T — the plate faces, not the thread. Left empty, the tool assumes 0.10, and says so.

Below it, the required clamp load. Two friction interfaces cut it in half. A twisting moment adds to it. An alternating shear needs a slip safety of 1.8, not 1.2. And a sealing pressure must be held as well.

Load introduction n says where the load enters the plates. Don't know? Leave it empty: the bolt and the clamp are each checked at their own worst case.

The required safety factor is your own target — shown, never deciding.

Step five: the checks. And here is our answer — FAIL, slip.

The bolt itself is fine: 91 % of yield after tightening, the threads don't strip, fatigue safety is over 20.

But look at the clamp. Minimum preload 18.5, minus the load 1.5, minus embedding 3.2 — leaves 13.8 kN per bolt.

Times four bolts, times μ_T 0.10: about 5.5 kN of friction against 7 kN of shear. Slip safety 0.79 — the bracket slips.

A label in the way? Drag it aside — its pointer stays on its part.

And what the tool does not check is listed too — off-centre clamping, hot joints, finite life — so silence never reads as a pass.

Step six: refinements. Limit p_G is the pressure your plate can take under the head — we entered 500 MPa. Without it, the verdict stays CHECK.

Washer, thread engagement and nut strength refine the pressure and stripping checks. Below: the tightening angle, and a thermal what-if.

So how do you fix it? The usual reflex: a bigger bolt.

M12 fixes the slip — but now the head crushes the plate: 621 against 500 MPa. FAIL, surface pressure.

Back to M10 — and tighten by angle instead of torque. The scatter drops to α_A 1.2: up to 31.3 kN, and 26 kN you can count on.

Slip safety 1.22. SAFE — same bolt, same plates.

But if those plate faces are oiled, μ_T drops to 0.07 — and it slips again. That one number decides the joint: measure it, or assume low.

One bolt is rarely alone. So the tool also gives you the force on every bolt — for any layout, and any load.

Choose bolt circle plus catalogue loads, and enter the loads the way a bearing catalogue states them: 24 M24 bolts on a 1000 mm circle — axial load 500 kN, tilting moment 800 kN·m, radial load 80 kN, and a torque of 10 kN·m.

Every bolt is coloured by its force — red in tension, blue in compression — and the most-loaded bolt is marked: 112.5 kN. Exactly the catalogue formula: four M_k over n D, minus F_a over n.

Loads can sit above the joint, too: the lever arm is drawn, with the tilting moment it creates.

Then one click runs the full VDI 2230 check on every bolt at its own load — and names the bolt that governs each check.

Every factor, formula and check goes into a PDF report your reviewer can follow.

That's VDI 2230. Try it on your own joint at mechanixcalc.com.