Bolts in Timber Connections

The wrong mental model

Bolts are the most familiar fastener in timber design. Anyone who has used a hex bolt anywhere - steel-to-steel, concrete anchor, machinery - already has a working mental model. The trouble is that mental model is usually wrong for timber, and the difference is not a small one.

A bolt in a timber connection does not work by friction. It does not develop clamping force. The threaded shank does not engage the timber as a screw would. What a bolt in timber actually does is sit in a slightly oversized hole and wait for the timber to push against it. Almost everything that matters about timber bolt design follows from that single fact.

What a bolt actually does in timber

When a steel bolt connects two steel plates with high-strength preload, it can carry load by friction across the joint surfaces - the bolt clamps the plates so hard that the friction can transmit load without the bolt itself bearing on anything. This is called a friction-grip or slip-critical connection, and it is the default model many engineers carry into timber from a steel background.

Timber does not work this way. The reason is simple - timber crushes plastically under the washer when you torque the bolt. Whatever clamping force you generate at installation, the timber yields locally over weeks and months until that clamping force is gone. There is no friction-grip equivalent in timber.

What you get instead is a bearing-driven connection. The bolt sits in a hole 1 to 2 millimetres larger than its diameter. Until enough load arrives to push the bolt against the hole edge, there is no resistance at all - the joint slides freely. Once bearing is engaged, the bolt transfers load by pressing on the timber, and the timber transfers it through embedment (crushing the local fibres around the bolt circumference). This is exactly the model behind Johansen's yield theory and EC5 §8.5's design tables.

The hole tolerance is not a tolerance in the precision sense - it is a fundamental feature of how timber bolts work. You cannot fit a bolt without it, and the joint slip it produces is a real component of the connection's behaviour, not an installation defect.

The yield moment that makes bolts ductile

The single equation most worth memorising for bolt design is the yield moment:

M_{y,Rk} = 0.3 · f_u · d^{2.6}

f_u is the bolt's ultimate tensile strength - typically 400 N/mm² for grade 4.6 mild steel, 800 N/mm² for grade 8.8 high-strength. The 2.6 exponent is empirical, fitted to test data, and matters more than its value suggests - doubling the diameter multiplies the yield moment by roughly six, not four as you might expect from a simple cube relationship. Bolt diameter is the single most powerful parameter you can change in a connection design.

This yield moment governs the Johansen mechanisms where the bolt forms plastic hinges. For slender bolts in dense timber, those plastic-hinge modes are the ductile ones - they absorb energy, deflect noticeably before failing, and give engineers the warning that the brittle splitting modes do not. Pushing your design into the slender-bolt regime is usually a sign of good detailing.

The Hankinson formula and the characteristic embedment strength equations apply identically to bolts as to other dowel-type fasteners - those are covered in the Timber Connection Design overview.

Spacing - what Table 8.4 says for bolts

The Eurocode 5 spacing rules for bolts are scaled to bolt diameter and load angle. The headline numbers for solid timber:

• a₁ parallel to grain: (4 + |cos α|) d
• a₂ perpendicular to grain: 4 d
• a₃,t loaded end distance: the larger of 7d or 80 mm
• a₃,c unloaded end distance: (1 + 6 sin α) d, never less than 4d
• a₄,t loaded edge distance: the larger of (2 + 2 sin α) d or 3d
• a₄,c unloaded edge distance: 3 d

These minima are larger than what you would specify for nails or screws of similar diameter, and the reason is bolt-specific. Bolt holes are pre-drilled, which makes splitting cracks easier to start because the wood fibres on the hole edge are cut clean rather than displaced as they are by a driven nail. The minima exist to keep splitting cracks from connecting between holes and propagating to a free edge.

For glulam and engineered timber the rules are similar but allow slightly tighter distances in some cases. The full table is in EC5 §8.5.1.1.

Washers - the rule that catches everyone

EC5 §8.5.2(3) specifies the minimum washer outer diameter:

d_w = min(12 · t_1, \; 4 · d)

For a 12 mm bolt in a 65 mm member, that is min(780, 48) = 48 mm. For a 16 mm bolt in the same member, min(780, 64) = 64 mm. The 4·d term governs in nearly all practical cases.

Why this matters: without a properly sized washer, the bolt head crushes a small area of timber under it as load develops. Once that local timber yields, the bolt loses its anchorage. The washer spreads the head load over enough area to keep the embedment stress below the timber's perpendicular-to-grain strength. In practice this is one of the most commonly missed details - designers specify the bolt, specify the spacing, and forget the washer.

Standard hardware-store washers are not always large enough. For an M16 bolt the standard washer is around 30 mm diameter, well below the 64 mm Eurocode minimum. Custom washers or plate washers are often required.

Rope effect - what bolts get that dowels don't

Bolts are unique among dowel-type fasteners in being able to carry significant axial load through the washer. EC5 §8.2.2(2) permits this rope-effect contribution, which adds to the lateral load capacity.

The rope effect is the axial resistance of the fastener combined with friction on the connected members as the fastener bends under lateral load. For bolts, the rope effect contribution is capped at 25% of the Johansen lateral capacity F_v,Rk. For dowels - which have no washer to anchor against - the rope effect is taken as zero.

The 25% cap matters in design because it can push a marginal connection over the line. If your Johansen analysis gives F_v,Rk = 16 kN per shear plane, the rope effect adds up to 4 kN, taking the design capacity to 20 kN. Ignoring the rope effect leaves usable capacity on the table.

For the rope effect to be valid, the washer has to be properly sized and the bolt installed with the washer fully engaging the timber. A loose bolt or one with an undersized washer gets no rope-effect contribution.

The pre-tensioning myth

A common mistake is to torque timber bolts at installation as if they were steel-to-steel. The intuition is that more torque means more clamping force means more resistance. This is not what happens.

When you torque a bolt against timber, you crush the wood under the washer. The clamping force you achieve at installation is real, but it is not stable. Over the following weeks and months, the timber yields under the load - partly elastically, partly plastically - and the clamping force decays toward zero. Within a year, the joint is essentially un-clamped regardless of how much torque you applied.

There is also no benefit to having clamping force in the first place. A bolt in a timber connection works by bearing on its hole edge, not by friction. The clamping force does not increase the joint's lateral capacity. The only thing it does is invite the timber-crushing failure mode under the washer that the spacing and washer rules are designed to prevent.

Standard guidance is to tighten the bolt firm-and-snug at installation - enough to seat the washer fully against the timber - and then revisit after twelve months. By then the timber has stabilised, the bolt has loosened as the wood dried, and a single re-tightening is enough for the connection's service life. Where re-tightening is not practical (concealed connections, high-rise frames), the design is done assuming reduced or absent clamping force from day one.

What goes wrong

The bolt failures engineers actually meet on site, in rough order of frequency.

End distance violations. The end distance gets compressed in detailing because the cleat or plate is short, and the splitting mode emerges first under load. Hard to spot in a single connection but obvious once a few in a structure have cracked.

Wrong washer size. Standard hardware washers used where Eurocode requires larger plate washers. The bolt looks correctly installed; the timber under the head is failing locally.

Over-tightening at installation. Crushes the timber under the washer immediately, and removes any rope-effect contribution before the structure is in service.

Ignoring moisture-driven loosening. Specifications without a 12-month re-tightening note assume the bolts stay tight forever. They do not.

Forgetting the perpendicular-to-grain strength is roughly fifty times lower than parallel. A bolt loaded across the grain near an edge can split the timber at a load far below the Johansen value because the splitting mode dominates the bearing mode.

Pulling it together

A bolt in timber is the simplest fastener to specify and the easiest to detail badly. The mechanism is bearing, not friction. The hole tolerance is fundamental. The washer is non-negotiable. The torque is a myth. The rope effect, properly anchored, gives you 25% of your lateral capacity back for free. Get these five things right and a bolt connection behaves almost exactly the way the Eurocode equations predict.