A request lands on the engineer’s desk: this enclosure needs to be strong, and it needs to keep moisture out. Picture the little computer that lives behind a roadside billboard — it just sits there taking calls and telling the sign what numbers to display, while the weather does its worst. What’s the best way to put it together?
The trouble is that “strong” and “sealed” aren’t specifications — they’re wishes. Strong against what: a drop, a pull, a vibration, someone prying it open? Sealed how well: splash-resistant, or submersible? Until those become real numbers, every joining method is equally defensible and equally likely to be wrong. The job of specifying a joint is mostly the job of defining what the joint actually has to survive.
This guide doesn’t reproduce historical WJT Associates material. It uses the same applied logic — define the requirement before you pick the method — to choose how to join a molded part without over-building it or under-sealing it.
First question: permanent or serviceable?
Before any method, one decision shapes all the others: does this assembly ever need to open again?
- Permanent. You put it together and it stays together — a one-time, sealed-for-life assembly. This opens the door to welding and bonding, which are strong and seal well but are difficult or destructive to reverse.
- Serviceable. It has to be opened for repair, battery changes, or rework — and still seal each time it’s closed. That points toward fasteners and gasketed joints that can be taken apart and reassembled.
Getting this backwards is expensive. Weld shut a box that needs servicing and every repair becomes a destroyed housing; use removable screws on something that should have been sealed for life and you’ve added cost and a leak path you didn’t need.
Second question: what is the joint actually fighting?
“Strong” has to be resolved into the specific things the joint will face, because different threats call for different methods:
- Mechanical load — pull-apart force, shear, a drop, sustained clamp load
- Vibration and fatigue — repeated cycling that loosens fasteners and cracks brittle joints
- Environmental sealing — splash, rain, humidity, dust, or full immersion
- Temperature — expansion mismatch, heat that softens adhesives or relaxes snap-fits
- Disassembly force — how hard a user (or a repair tech) is meant to be able to open it
Write those down as the load case before choosing a method, the same way a usable acceptance standard names its conditions instead of saying “looks good.” “Survives a 1-meter drop and resists splashing from any angle” is a spec. “Strong and sealed” is an argument waiting to happen.
The methods, and what each is good at
No joining method is best in the abstract — each trades strength, sealing, serviceability, and cost differently. The realistic shortlist for molded parts:
| Method | Permanent or serviceable | Sealing | Notes |
|---|---|---|---|
| Screws into bosses | Serviceable | With a gasket/seal | Familiar, reworkable; bosses need good design or they strip/crack |
| Snap-fits | Usually serviceable (some permanent) | Poor by itself | Lowest piece cost, no fasteners; needs careful geometry and draft; can relax over time/heat |
| Press fits | Semi-permanent | Limited | Simple; relies on interference and material creep behavior |
| Ultrasonic welding | Permanent | Good with proper joint | Fast, strong, no consumables; needs a designed energy-director joint and compatible resins |
| Heat staking | Permanent | Localized | Good for capturing/attaching components; not a full seal by itself |
| Solvent or adhesive bonding | Permanent | Good to excellent | Strong and sealing; cure time, surface prep, and chemical compatibility matter |
| Gaskets / O-rings | Serviceable (with fasteners) | Excellent | The sealing element, not the structure — pair with screws or snaps |
A few realities behind the table. Snap-fits are the cheapest joint per part and need no hardware, but they don’t seal on their own and their performance lives or dies on geometry, draft, and how the material relaxes over time and temperature. Ultrasonic welding is fast, strong, and consumable-free, but it demands a properly designed joint (an energy director) and compatible materials — it’s a design decision made at the part stage, not a fix added later. And a gasket or O-ring is a sealing element, not a structure: it makes a serviceable joint watertight, but something else (usually screws) still has to hold the load.
Match the method to the requirement
Put the two questions and the table together and the choice usually falls out:
- Define serviceability. Permanent unlocks welding and bonding; serviceable points to fasteners and gaskets.
- Define the load case and the seal level as real, testable requirements — drop height, pull force, splash versus immersion, temperature range.
- Shortlist methods that meet both, then choose on cost and manufacturability.
- Design the joint, not just specify it. Bosses, snap geometry, weld energy directors, and gasket grooves are part features that have to be designed in from the start — draft, wall thickness, and material all interact with the joint.
The discipline is the same one that runs through good part standards and good troubleshooting: resolve the vague word into a measurable requirement first. “Strong” and “sealed” can’t be engineered. “Holds 200 N of pull-apart, survives a 1-meter drop, and keeps out driven rain across its temperature range” can.
FAQs
How do I decide between a snap-fit, screws, or welding?
Start with one question: does the assembly ever need to open again? If it’s permanent, welding (or bonding) gives you a strong, well-sealing joint with no hardware. If it needs servicing, screws into bosses — usually with a gasket for sealing — let you open and reclose it. Snap-fits are the lowest-cost option and need no fasteners, but they don’t seal by themselves and depend heavily on geometry. After serviceability, choose on the actual load and seal requirements, then on cost and manufacturability.
Why aren’t “strong” and “sealed” usable specifications?
Because they don’t say strong against what or sealed to what level. “Strong” could mean resisting a pull-apart force, surviving a drop, withstanding vibration, or resisting prying — each pointing to a different joint. “Sealed” could mean splash-resistant or fully submersible, which are very different requirements. Until they’re written as testable numbers — pull force, drop height, splash versus immersion, temperature range — any method looks defensible and you’re likely to over-build the strength while under-sealing, or vice versa.
Does a gasket or O-ring make a joint strong?
No. A gasket or O-ring is a sealing element, not a structural one. It makes a serviceable joint watertight, but something else — typically screws into bosses — still has to carry the mechanical load and provide the clamping force that compresses the seal. Specifying a gasket without specifying the fastening that holds it is only half a joint. Think of the gasket as the seal and the fasteners as the structure; a sealed, serviceable assembly needs both.
What’s the difference between a permanent and a serviceable joint?
A permanent joint is assembled once and stays closed for life — welding, bonding, and some snap-fits — which generally gives strong, reliable sealing but is difficult or destructive to reverse. A serviceable joint is designed to be opened and reclosed for repair or maintenance, and still seal each time, which points to fasteners with gaskets. Choosing the wrong one is costly: welding a housing that needs servicing destroys it on every repair, while using removable fasteners on a sealed-for-life part adds cost and an unnecessary leak path.