On the kart: every bolt, from the spark plug (140-200 lb-in, by the Briggs rules) to the steering wheel (minimum 3 bolts, by the CCKRA fastener rule), the caliper bolts with their safety pins, the header bolts with their twisted wire, the nylock nuts on the tie rod ends. Tech inspection is mostly a walk down this list.
The principle in one sentence
A tightened bolt is a stretched spring, and it is the stretch, not the torque, that holds the joint together.
The one equation
preload (lb) = torque (lb-in) / (K x bolt diameter (in))K is a friction factor, about 0.2 for a dry, plain bolt (typical). The preload is the tension the bolt is left holding after you stop turning the wrench. Torque is how you put it there; preload is what does the work. Most of the torque (roughly 85-90%) is spent overcoming friction in the threads and under the head; only the remainder becomes stretch. That's why K matters so much and why an oiled bolt at the same torque ends up far tighter than a dry one.
Math: torque units. Torque is force times arm (M01). lb-in means pounds times inches; lb-ft means pounds times feet. 1 lb-ft = 12 lb-in, so 140 lb-in = 11.7 lb-ft. The metric unit is the newton- meter: 1 lb-in = 0.113 N-m, 1 lb-ft = 1.356 N-m. Torque wrenches come in all three; the spark plug spec of 140-200 lb-in is 11.7-16.7 lb-ft or 15.8-22.6 N-m. Getting this wrong by a factor of 12 strips threads, and it's the commonest wrench mistake there is.
Why it works
A steel bolt is stiff, but not rigid. Tighten it and it stretches a few thousandths of an inch, like a very strong spring pulled slightly open, and it pulls the two parts together with the spring's force. That clamping force is what stops the parts sliding, rattling and working loose. As long as the load trying to separate the parts is smaller than the preload, the parts never move and the bolt never feels the load change: the joint behaves like one solid piece.
A loose bolt is a bolt with no stretch. Now every bump lets the parts shift a little, the bolt sees the load as a repeated tug, and steel that's tugged repeatedly cracks (fatigue). Loose bolts don't fail because they fell out; they fail because they broke first. The fixes in the rule book are backups against the first symptom, not the cause:
- Nylock nut: a nylon ring grips the thread so the nut can't spin off if preload is lost. It does not restore preload.
- Safety wire / pin: a physical stop against rotation. Same.
- Double nuts (ballast rule): the second nut jams the first.
All three keep a loose bolt from leaving. None of them make a loose bolt tight. Torque does that, once, correctly.
On our kart, in numbers
The spark plug. Briggs specifies 140-200 lb-in and an unaltered plug with its sealing washer. The washer crushes to seal the combustion chamber; the torque range is what crushes it correctly without stripping the aluminum head. Under-torque: a leak and a plug that backs out. Over-torque: threads pulled out of a sealed engine you're not allowed to repair. The number is not a feel.
A 5/16 bolt. At 150 lb-in with K = 0.2: preload = 150 / (0.2 x 0.3125) = 2400 lb. That is the force one small bolt holds a ballast plate or a seat strut with. Its stretch: a steel bolt with 0.052 in^2 of thread area, 1.5 in of grip, stretches by (force x length) / (stiffness x area) = 2400 x 1.5 / (30,000,000 x 0.052) = 0.0023 in, about two thousandths. (30,000,000 psi is steel's stiffness, the 200 GPa of M09 in inch units.) Invisible, and it's everything.
The fastener audit. The CCKRA list (pedals, brake rod and its tether, master cylinder, calipers, kingpins, tie rod ends, steering hub, steering wheel with 3 bolts minimum, third bearing with 2) is every joint whose loss ends the session or worse. Each one needs preload AND a backup. The 9/19 main's "scraping something" hasn't been found yet; a part that lost its preload is the first place to look, because that's how most scrapes start.
Two things to notice
- Torque wrenches measure the wrong thing, but it's the best we've got. They read the torque, and the preload depends on K, which depends on whether the thread is dry, oily, dirty or new. Same torque, different bolt condition, different clamping. Consistency (same condition every time) matters more than the exact number.
- Re-torquing after heat is a real thing. Aluminum expands more than steel; a hot engine's bolts see their preload change, and exhaust studs in particular loosen with cycles. The Briggs rule that header bolts must be wired is an admission of this.
Problems
1. Convert the plug spec. 140-200 lb-in to lb-ft, and to N-m. If the only torque wrench in the kit reads lb-ft and clicks in whole numbers, what setting do you use?
2. Oil on the threads. An oiled bolt has K about 0.15 instead of 0.2 (typical). Same 150 lb-in on the 5/16 bolt: what is the preload now, and by what percentage did it rise? Is that toward stripping or toward loosening?
3. Which bolt matters most? From the fastener-audit list, pick the one whose failure is worst, and the one most likely to loosen (hint: the ones that see heat and vibration together). Are they the same bolt? What backup does each have on our kart?
Go look: put a hand on every bolt in the audit list and try to turn the nut with your fingers. Find the nylon ring in a nylock (it's visible at the top of the nut). Follow the safety wire on the header and see that it's twisted the way that would tighten, not loosen, the bolt if it moved. Then compare with the August photo of the header wire in the kart's records and confirm it's the same wire.