Fastener Specifications & Technical Data

A #8 screw is 0.164 inches thick. An 8d nail might be 0.131 or 0.113 depending on which bin you grabbed from. A 3/8 lag takes a 9/16 socket. None of that is memorable, all of it is the kind of thing you need halfway through a cut, and most of it is buried in standards that cost money to read. This page is the lookup layer: dimensions, sizes and selection guidance for the fasteners a woodworker actually buys.

What size is a #8 screw, really

Screw gauge is a diameter, and it is the one number on the box that tells you something physical about the fastener. It is not a length, it is not a strength rating, and it is not linear: the step from #6 to #8 is 0.026 inches, and the step from #14 to #16 is the same 0.026, but the second one is a much smaller proportional jump.

Three diameters get quoted for the same screw, and catalogues pick whichever one flatters the product. The shank diameter D is measured across the threads and it is what governs holding power. The root diameter Dr is the solid core between the threads, and it is the cross-section that actually carries load and the one that snaps. The head diameter DH decides how big a countersink you need.

Odd gauges stop at #9. There is no #11, #13 or #15 wood screw in the standard, and above #10 the sequence runs even only. If a supplier offers you a #11, it is not a standard size.

Head styles, and what each one is for

The head does two jobs: it transfers load into the wood surface, and it decides whether the screw finishes flush, proud, or buried. Only one common head is designed to end up level with the surface, and it needs a countersink cut to match.

American flat heads use an 82 degree included angle. Metric countersunk screws use 90 degrees. They look interchangeable in a bin and they are not: an 82 degree screw in a 90 degree hole bears on the rim of the hole rather than on the full cone, which concentrates all the load on a thin ring of fibre and is a reliable way to split thin stock on the last quarter turn.

Drive types and why Phillips slips

Cam-out is the bit climbing out of the recess under load. In a Phillips drive it is not a defect, it is geometry: the flanks are tapered, so torque produces a force component along the axis of the driver that pushes it back out. On a 1930s assembly line with no clutch that was useful, because camming out was better than shearing the head off. In a cordless drill with a clutch it just destroys the recess.

Square and Torx recesses have parallel walls. The driving force is transmitted sideways with nothing trying to eject the bit, which is why they tolerate far more torque before anything rounds off. Torx spreads that force over six lobes instead of four flats, so it goes further still. This is why essentially every structural screw sold today ships with a Torx or square drive.

Coarse, fine, and the thread that will not close a joint

Thread form is chosen against wood density. Coarse threads are deeper and further apart, which cuts a bigger anchor in low-density softwood and sheet goods. Fine threads remove less material per turn, which matters in hardwood where a coarse thread acts like a wedge and splits along the grain.

But the distinction that actually ruins joints is not coarse versus fine. It is partial versus full thread.

A partially threaded screw has a smooth shank under the head. That shank spins freely in the top board, so the threads in the lower board can drag the two pieces together. A fully threaded screw bites both boards at once, and whatever gap exists at the moment the threads engage is the gap you are left with. Driving harder does not close it. It strips the pilot hole, and then the screw does nothing at all.

Two fixes. Use a partially threaded screw with the shank at least as long as the top board. Or drill the top board out to full clearance, so the threads cannot grip it no matter how deep they run.

How long a screw do you actually need

The standard sets a minimum. NDS 12.1.5.6 puts minimum penetration into the receiving member at six shank diameters, and 12.1.6.4 says the same for nails. That measurement includes the tapered tip, which matters more than it sounds: the tip is roughly two diameters of the screw that goes into the wood without carrying full-depth thread.

Treat those as a floor rather than a target. A screw that only just clears 6D is legal and barely holding. The number worth thinking about is how much full-depth thread ends up in the receiving board, which is the screw length minus the top board minus the tip.

An 8d is three different nails

The penny system is a survival from when nails were priced by the hundred. What it fixes today is roughly the length. It does not fix the diameter, and diameter is what determines both how much the nail holds and how likely it is to split the wood.

A common nail is the thick one. A box nail of the same penny size is noticeably thinner, which is the entire reason it exists: less steel displaces less wood and splits less, at the cost of holding less. A sinker is thinner again and also shorter, with a coated shank and a countersunk head meant to be driven flush.

The last column is the one to read. If a plan calls for 8d common and you use whatever the store labelled 8d, you can lose about a quarter of the fastener's cross-section without anything looking different.

Shank profile decides whether the nail stays put. A smooth shank resists withdrawal by friction alone, and friction falls away as the wood dries and shrinks back from the steel. That is the mechanism behind a deck that grows popped nail heads in its second summer. Ring shanks trade drivability and any hope of removal for a large gain in withdrawal resistance.

Lag screws and the socket that fits

A lag screw is a heavy wood screw with a hex head, driven with a wrench rather than a driver. The dimension people need mid-job, usually with the lag already started and the socket set open on the floor, is the width across flats.

It is not intuitive, because the head grows faster than the screw. A 1/4 inch lag takes a 7/16 socket. A 1/2 inch lag takes 3/4. The ratio drifts the whole way up the range, so there is no rule of thumb to fall back on.

The unthreaded shank length matters too. The clearance hole in the piece being attached has to run at least as deep as that smooth section, or the shank binds in the top board and the joint never pulls closed, exactly as with a fully threaded screw.

Bolts, and why a washer is not optional

A bolt is a different mechanism from everything above. A screw holds because threads grip wood fibre. A bolt passes through clearance holes in both pieces and clamps them, so the load rides on a steel shank in shear plus friction between the clamped faces.

Root diameter here is based on the UNC coarse thread series. A fine-thread bolt of the same nominal size has a shallower thread and therefore a larger root, which is why fine threads are stronger in tension and worse at tolerating a damaged thread.

Steel does not much care whether there is a washer under the nut, which is why machine practice often skips them. Wood cares enormously. Wood crushes perpendicular to the grain at a small fraction of the stress steel tolerates, so a bare nut face dishes into the surface, the joint loses its clamp as soon as the fibre yields, and re-tightening only digs the dish deeper.

Which coating, and when it is a code requirement

Most of the time coating choice is a judgement call about how wet the thing will get. In one case it is not a judgement call at all.

Modern pressure-treated lumber uses copper-based preservatives, ACQ and CA-C, which carry substantially more copper than the arsenical treatments they replaced. Copper in wet contact with plain zinc plating consumes it. IRC R317.3 therefore requires fasteners in contact with preservative-treated wood to be hot-dip galvanized, stainless steel, silicon bronze or copper. This is the one place on this page where guessing has a code consequence.

Galvanic corrosion, and the pairing everyone gets backwards

Put two dissimilar metals in contact with moisture between them and you have built a battery. One of them is consumed to protect the other. Which one depends on where they sit relative to each other in the galvanic series.

Two consequences catch people out. Galvanized fasteners against aluminum flashing sit close together near the active end, and the zinc coating goes first, quickly, because a thin coating has very little material to give up. And stainless against copper corrodes the copper. Stainless is more noble than copper, not less, which is the reverse of the ordering shown on a lot of builder-facing charts. If you are running stainless screws through copper flashing, the flashing is the part that fails.

Why the same screw holds differently in pine and oak

Withdrawal resistance scales with wood density, and it scales faster than linearly. That is why a screw that feels solid in Douglas fir feels loose in white pine and snaps in white oak before it seats.

The practical reading is at the two ends. In low-density softwood the failure mode is the threads stripping their own hole, so you want coarse threads, more fasteners, and no more torque than it takes to seat the head. In dense hardwood the wood will not yield, so the fastener does: correct pilot and clearance holes stop being optional, and a fine thread is worth the trip to the hardware store.

Where this page hands off

This page stops at dimensions and selection. Three siblings pick up where it ends.

• Pilot hole chart for the drill schedule by screw gauge and species, and for how much of your screw is actually engaged.
• Nails vs. screws for failure modes, and for why a toenailed connection is derated.
• Structural lags and bolts for installation, spacing, edge and end distances, and deck ledger schedules.

For load-carrying capacity rather than geometry, our standards reference holds the NDS design values with their citations, and assumptions and methodology explains how we use them.

Common questions