Technical note 4

TN-4: Fastening Systems Survey

Which fastener holds the chair for a century, and how much of the answer we actually know. First draft 19 August 2026, from KU's questions about the screw versus the preloaded bolt. Companion to TN-7 (Why Joints Go Slack), which owns preload and creep; this note owns the fastener itself. Status: DRAFT; the history and corrosion claims are VERIFIED as marked, the mechanics are STANDARD, and the recommendation is DERIVED and, where it is a guess, says so.

The claim

The best joint carries no load in its fastener at all: large faces abutting in compression, low moment, and a fastener whose only job is to position the members and keep the joint closed. The Keel Chair's truss architecture is good this way and not perfect, so some fastener load remains. Two honest findings follow. First, the longest-PROVEN joint of exactly this kind uses no metal at all: the drawbored pegged mortise and tenon, a wooden peg through offset holes that pulls the shoulders tight and holds them in compression, with millennia of record. Second, for the metal fasteners we are actually choosing between, the evidence is uneven, and the specific system we had defaulted to, a preloaded sleeve and bolt with a spring washer, is the one with the least record. The frame solve tells us where each belongs.

What we actually know, ranked by record

So is a legible screw better than a preloaded bolt? Partly a guess.

We are not guessing blindly. We reason from four bases that do have records: MECHANICS, a fastener that only positions a compression joint beats one carrying withdrawal or shear, and the Keel Chair already minimizes fastener load; CORROSION, bronze lasts and iron does not, worst in oak, favoring bronze whatever the fastener's form; HISTORICAL RECORD, pegs and bronze for millennia, iron screws for corroding centuries, the bolt-sleeve-spring for nothing; and REPAIRABILITY, where the screw wins decisively, because KU's repair ladder is exactly standard conservation practice and is legible, nut-free, and tool-simple: a bigger screw, then dowel and redrill the hole, then a filler, then a threaded insert, then a through-bolt, and finally replace the wood.

Put those together and the evidence points somewhere more specific than "bolt good." A LEGIBLE BRONZE SCREW in a compression joint may be the better longevity-maximizing fastener for most of the chair, because it stacks bronze's corrosion record onto the screw's three-century repair tradition, and in a compression joint it carries almost no load anyway. The preloaded bolt-sleeve-spring earns its complexity only where a joint must resist working loose under seasonal cycling, which the frame solve identifies as the TIE, the one member in tension (TN-7's preload argument applies there and only there). So the doctrine is: match the fastener to the joint. Bronze screws in the compression joints; the engineered preload reserved for the tie; and the bolt-sleeve-spring flagged honestly as the option we chose by reasoning, not by record.

Are we essentially guessing?

On the bolt-sleeve-spring specifically, largely yes, and the note should be humble about it. We have strong records for pegs, bronze, and screws, and no record for that synthesis. The conservative reading is to prefer the proven, a bronze fastener in a compression joint, or a drawbored peg where the member is solid wood, and to treat the preloaded bolt as an engineered exception for the tie, carrying an explicit "unproven at century scale" flag until something better than reasoning supports it.

For the Standard

A fastener clause. Design every joint so its faces abut in compression and the fastener only positions. Match the fastener to the joint's load, using the frame solve: a legible bronze screw at the compression joints; a preloaded bronze bolt with a movement-taking washer at the tie, the one tension member. Prefer copper alloys (silicon bronze) over ferrous fasteners everywhere, and never ferrous in oak. Document the repair ladder with the chair so a future custodian has it. Where a member is solid wood rather than ply, a drawbored peg is a proven fastener-free alternative worth offering.

Open questions

  1. Screw withdrawal and shear in the actual materials: structural plywood (thin plies, end-grain-into-ply is weak) versus the solid hardwood keel. Measure, do not assume.
  2. The bolt-sleeve-spring has no accelerated-aging validation in this project. A cyclic-humidity test of a preloaded tie joint, tracking clamp force over cycles, would move it from guess toward evidence.
  3. Drawbore in ply is a poor fit (ply splits, plies shear); the peg option is for the solid keel and any future solid-wood variant, not the panels.
  4. Glue interacts with all of this and is its own question, deferred to a TN-2 (Adhesives and Reversibility): whether a fastened compression joint should be glued at all, and if so with what.

Revisions

19 Aug 2026: first draft, from KU's fastener questions. The central honest finding, that the drawbored peg and the bronze screw are proven while the preloaded bolt-sleeve-spring is not, and the frame-solve rule for where each belongs, both date from this pass.

Renumbered 19 Aug 2026: first written the same day as TN-13 (Fasteners), then folded into its reserved slot, TN-4 (Fastening Systems Survey), to retire the duplicate. The content is unchanged; the survey of barrel nuts, cross dowels, and the Lamello P-System that this slot reserved will fold in on the next pass.

Technical Notes of the Hundred-Year Chair: public from first draft, refined iteratively, forever. A claim is VERIFIED, STANDARD, DERIVED, or UNVERIFIED, and says which. Sources: the WAG/AIC wood-screw study, the drawbore and joinery-history accounts, and the oak-iron corrosion record (ledger item 36). The register · the verification ledger · hundredyearchair.com