Mechanical Fastening
Mechanical fastening joins parts with screws, inserts, clinch hardware or snap fits so the joint can be assembled and taken apart again.
- Part
- Joining
- Revised
- 2026-08-11
At a glance
- Family
- Mechanical
- Typical tolerances
- Normal-fit clearance holes about 1/32 in (0.8 mm) over nominal per ASME B18.2.8 for screws up to 1/2 in; preload scatter ±25–35% with torque control, ±15% with angle control, ±5–10% with bolt elongation measurement
- Typical volumes
- 1 to 10,000,000+
- Lead time
- Off-the-shelf hardware in hours to days; seconds per fastener at assembly
- Materials
- Metal, Plastic, Wood, Composite
What it is
Mechanical fastening joins parts with screws, bolts, threaded inserts, self-clinching hardware or molded snap fits, so the assembly can be taken apart again without destroying it. That serviceability is the reason it survives against faster and cheaper joining processes.
A bolted joint does not work the way most drawings imply. Clamp load, not the bolt shank, carries shear in a properly designed joint — friction between the clamped faces does the work, and the bolt only sees shear in a bearing-type connection. Preload is therefore the design variable, typically targeted at 65–75% of the fastener's proof load.
Getting that preload is the hard part. Torque control delivers it with ±25–35% scatter because most of the applied torque is consumed by friction; angle control tightens that to about ±15%, and direct bolt elongation measurement to ±5–10%. Volumes run from one-off to tens of millions of joints a year.
How it works
- Select the fastener and the thread engagement. Engagement length is set by the weaker material: about 1x nominal diameter into steel of similar strength, 1.5–2x into aluminum, 2x into magnesium, and 2–2.5x into thermoplastic. Too little engagement strips the female thread before the bolt reaches preload.
- Prepare the hole. Tapped holes are drilled to the tap drill size for the required thread engagement percentage — see the tap drill chart and the drill size chart. Clearance holes follow ASME B18.2.8: a normal-fit clearance hole is about 1/32 in (0.8 mm) over nominal for screws up to 1/2 in. Thread dimensions and pitch are on the thread size chart.
- Tighten and develop preload. Torque relates to preload through T = K x D x F, where K is the nut factor — about 0.20 for plain steel, 0.15 for zinc-plated or lubricated, 0.10 with wax or MoS2. Because K varies with surface condition, so does the resulting clamp load: this is the source of the ±25–35% scatter. Torque values by size, grade and lubrication are on the bolt torque chart.
- Verify. Residual or breakaway torque audits, torque-angle signature monitoring on production tools, or ultrasonic bolt elongation measurement where the joint is critical.
The other mechanical fastening routes
Threaded inserts in plastic. Heat-set or ultrasonic inserts give the highest pull-out and torque-out in thermoplastic because the melt flows into the knurl. Design the boss outside diameter at roughly 2x the insert outside diameter.
Self-tapping screws in plastic. Thread-forming screws for ductile thermoplastics (they displace rather than cut material, so there are no chips); thread-cutting screws for brittle and filled resins and for thermosets.
Self-clinching hardware. Nuts, studs and standoffs pressed into sheet metal, providing a load-bearing thread in material far too thin to tap. They require sheet softer than the fastener — typically HRB 80 or below for standard steel clinch hardware.
Snap fits. A molded cantilever or annular feature that deflects during assembly and returns to retain. No hardware at all, but every snap is a strain event in the resin.
Design guidelines
Design the joint around preload, not around torque
Target 65–75% of proof load for a joint that will be reused, and understand that torque control alone delivers that with ±25–35% scatter. If the joint's function depends on clamp load — a gasketed face, a friction-critical connection, a fatigue-loaded bolt — either specify angle control (±15%) or design so the scatter does not matter. Preload also has to survive: use a hardened washer to prevent embedment, and keep soft or gasket material out of the clamp path where it will creep and shed preload.
Get the thread engagement right
Engagement is set by the weaker material, not by the bolt: 1x diameter into steel, 1.5–2x into aluminum, 2x into magnesium, 2–2.5x into thermoplastic. In a thin plate that cannot provide it, use a self-clinching nut, a threaded insert or a nut plate — not a longer screw into more of the same thin material.
Bosses and inserts in plastic
Boss outside diameter should be about 2x the insert or screw major diameter. Add a 0.010–0.015 in (0.25–0.4 mm) counterbore at the top of the boss so the first thread does not blow it open. For self-tapping screws, the ratio of stripping torque to driving torque should be at least 3:1, and 5:1 is a better target for automated assembly — below 3:1 the driver will strip bosses routinely.
Snap fits
Design a cantilever snap so maximum bending strain during assembly stays within the resin's published allowable — commonly around 2–4% for unfilled engineering thermoplastics on a one-time assembly, roughly 60% of that for a joint that cycles repeatedly, and 1–2% or less for glass-filled grades. Always take the allowable from the specific resin datasheet, and taper the beam so strain is distributed rather than concentrated at the root.
Fastener and substrate compatibility
| Pair | Notes |
|---|---|
| Steel screw into steel | Baseline; 1x diameter engagement |
| Steel or stainless screw into aluminum | 1.5–2x engagement; galvanic couple — isolate or seal in wet environments |
| Stainless screw into stainless | Prone to galling under load; use a dissimilar grade, a lubricant or an anti-seize |
| Screw into magnesium | 2x engagement; strongly anodic — isolation is mandatory |
| Screw into thermoplastic | 2–2.5x engagement; use thread-forming screws or a heat-set insert |
| Screw into thermoset or 30%+ glass filled | Thread-cutting screws; thread-forming will crack the boss |
| Clinch nut into sheet | Sheet must be softer than the fastener, typically HRB 80 or below |
| Fastener into composite | Bearing failure governs — use a washer or a bushing to spread load |
For dissimilar-metal joints, keep the anodic index difference under about 0.25 V in harsh environments and 0.50 V in benign ones, per MIL-STD-889, or isolate with a washer, coating or sealant.
Inspection
Torque audits — residual or breakaway — are the standard production check, though they measure torque rather than the preload that actually matters. Torque-angle signature monitoring on DC electric tools catches cross-threads, missing washers and soft joints in real time and is the practical high-volume method. For critical joints, ultrasonic bolt elongation measures preload directly. Threaded inserts and self-clinching hardware are verified by pull-out and torque-out testing on lot samples.
| Feature | Recommended | Limit | Why |
|---|---|---|---|
| Preload target | 65–75% of proof load | Reusable joints | Leaves margin against yield and relaxation |
| Preload scatter, torque control | Expect ±25–35% | ±15% with angle control | Friction consumes most of the applied torque |
| Thread engagement into aluminum | 2x nominal diameter | 1.5x minimum | Female thread strips before the bolt reaches preload |
| Boss outside diameter for an insert | 2x insert diameter | — | Hoop strength around the knurl |
| Strip-to-drive torque ratio | 5:1 | 3:1 minimum | Below 3:1 assembly strips bosses routinely |
| Snap fit assembly strain | Per resin datasheet | ~2–4% unfilled, 1–2% filled | Exceeding it cracks the beam on first assembly |
| Dissimilar metal anodic index difference | Under 0.25 V harsh | 0.50 V benign | Galvanic corrosion of the anodic member |
Cost drivers
Fastened assembly cost lives in three places: the purchased hardware, the features that receive it, and the labor to drive it. The hardware is usually the smallest of the three. Every fastener also implies a hole, a boss or a tapped feature in both parts, and a driving operation with its own station, tool and verification step.
Part count is therefore the lever. A design with twenty screws in five sizes needs five bit changes or five stations; the same design with twelve screws in one size needs one. Unique fastener part numbers carry their own cost in purchasing, inventory and line-side presentation.
Torque verification is a genuine recurring cost on anything safety-related — logged DC tools, audit sampling and the documentation behind them. That cost scales with the number of joints that need verifying, not with the number of fasteners.
- Standardize on one or two fastener sizes per assembly and one drive type. Bit changes and tool stations cost more than the hardware.
- Delete fasteners with snap fits or staking on non-serviceable joints — a snap fit costs nothing per unit once the mold exists.
- Design captive hardware in. Self-clinching nuts, nut plates and captive screws remove the loose-part handling that dominates manual assembly time.
- Do not tap thin sheet. A clinch nut is cheaper than the extra material or the second operation needed to get thread engagement.
- Reserve torque verification for joints that need it. Blanket logged-torque requirements are a common and expensive overspec.
Questions
5 questionsWhat preload should a bolted joint be designed for?
65–75% of the fastener's proof load for a joint that will be taken apart and reused. That leaves margin against yield and against the relaxation that follows initial tightening. Remember that torque control delivers this with ±25–35% scatter, because most of the applied torque is spent overcoming thread and underhead friction rather than stretching the bolt.
How much thread engagement do I need in aluminum?
1.5–2x the nominal fastener diameter — so a 1/4-20 screw needs roughly 0.375–0.500 in (9.5–12.7 mm) of engaged thread. Engagement is set by the weaker material: 1x diameter suffices in steel of comparable strength, 2x is needed in magnesium, and 2–2.5x in thermoplastic. If the part is too thin to provide it, use a clinch nut or a threaded insert.
Why does torque control give such variable clamp load?
Because roughly 85–90% of applied torque is consumed by friction under the head and in the threads, and only the remainder actually stretches the bolt. Any variation in surface finish, plating, lubrication or debris changes the nut factor K in T = K x D x F, and with it the preload. That is why torque control scatters ±25–35% while angle control reaches about ±15%.
How do I keep a self-tapping screw from stripping a plastic boss?
Design for a stripping-torque to driving-torque ratio of at least 3:1, and 5:1 for automated assembly. That comes from correct pilot hole diameter for the resin, a boss outside diameter of about 2x the screw major diameter, and a 0.010–0.015 in (0.25–0.4 mm) counterbore at the top of the boss so the first thread does not split it. For joints that will be opened repeatedly, use a heat-set insert instead.
When should I use a self-clinching nut instead of a tapped hole?
Whenever the sheet is too thin to give the required thread engagement — which is most sheet metal below about 0.090 in (2.3 mm). A clinch nut presses permanently into the sheet and provides full-strength threads without a second operation or added material. The sheet must be softer than the fastener, typically HRB 80 or below for standard steel clinch hardware.