Assembly Design: Screws, Snaps, Ultrasonics and Adhesive Compared

Four ways to join two parts

Almost every moulded assembly is held together by one of four methods: a screw, a snap fit, an ultrasonic weld or an adhesive. They differ in cost, in strength, in sealing and, most importantly, in whether the joint can be opened again. That last question is usually decided by the service strategy rather than by the engineer.

A screw is a mechanical fastener that clamps two parts together. It is adjustable, it can be removed, and it needs a boss in one part and a clearance hole in the other. A snap fit is a moulded feature that deflects and clicks into place. It is fast, it needs no hardware, and it is permanent in the sense that it can be opened a few times and then not.

An ultrasonic weld melts the two parts together at a joint line. It produces a joint as strong as the material, it seals, and it is completely permanent. An adhesive bonds the surfaces with a third material, which can seal and can accommodate dissimilar substrates, at the cost of surface preparation and cure time.

The choice is usually made early, on the basis of one criterion, and then defended for the rest of the project. A structured comparison at the concept stage avoids the situation where a snap fit is chosen for cost and then a service requirement appears that only a screw can meet.

The cost model

Each method has a fixed cost and a variable cost, and they are structured differently. A screw has almost no fixed cost and a high variable cost: the hardware plus the labour to drive it. A snap fit has a fixed cost in the tool and almost no variable cost, because the joint forms itself when the parts are pressed together.

An ultrasonic weld has a large fixed cost in the machine and the tooling, and a moderate variable cost in cycle time. An adhesive has a moderate fixed cost in dispensing equipment and a low variable cost, plus a hidden cost in surface preparation and in the fixtures that hold the parts while the adhesive cures.

The chart on the left is a real comparison for a small housing: screws at EUR 0.42 each plus EUR 2,000 of fixtures, snap fits at EUR 0.06 plus EUR 8,000 of tool modification, ultrasonic at EUR 0.18 plus EUR 45,000 of machine and tooling, and adhesive at EUR 0.12 plus EUR 12,000 of dispensing and fixtures.

The result is that snap fits win at every volume in this case, because their variable cost is so much lower and their fixed cost is modest. That is not a general conclusion. On a part that needs to be sealed, or on a part with a wall too thin to carry a snap, the ranking changes completely.

Screws: adjustable and slow

A screw is the only one of the four that can be undone and redone without damaging anything. That property is worth a great deal in products that are serviced, in prototypes, and in anything where the assembly might have to be opened during commissioning.

The cost is in the labour and in the boss. A screw needs a boss with enough wall to take the thread, a clearance hole in the mating part, and enough depth for the thread engagement, which is typically one and a half times the diameter in plastic. Four screws on a small housing add 8 g of material and 30 seconds of assembly.

Screws also introduce a stress concentration. A boss that is too thin cracks when the screw is driven, and a boss that is too thick sinks on the opposite face. The design rules for a screw boss are specific, and they are the reason a screw is not as simple as it looks.

Thread forming screws for plastic are the usual choice, because they cut their own thread and tolerate a wider boss tolerance than a machine screw with an insert. A metal insert is better for anything that will be assembled more than twice, at a cost of about EUR 0.15 per insert plus the moulding step.

Snap fits: fast and unforgiving

A snap fit is a cantilever or a ring that deflects during assembly and returns to its original shape behind an undercut. It costs nothing per part, it assembles in under a second, and it needs no hardware. It is the cheapest joint available and the one with the least tolerance for error.

The design rules are strict. The strain at the root of the cantilever must stay below the allowable strain for the material, which for an unfilled amorphous plastic is about 2 percent at 23 C and much less at low temperature. A snap that is designed at 3 percent strain will work in the lab and crack in a cold warehouse.

The retention force and the assembly force are set by the same geometry, which is the fundamental compromise. A snap that is easy to assemble is easy to pull apart; a snap that holds firmly is hard to press together. The usual solution is an asymmetric hook, with a shallow lead-in angle and a steep retention angle.

Snap fits are also the joint most sensitive to moulding variation. A 0.1 mm change in the undercut depth, which is within normal moulding tolerance, changes the retention force by 20 percent. That is why snap fits are usually developed with a tool that has adjustment, or with a first article that is measured rather than just fitted.

Ultrasonic welding: permanent and fast

Ultrasonic welding uses high frequency vibration to melt a small triangular feature, called an energy director, at the joint line. The melted material flows and bonds, and the result is a joint that is typically 80 to 100 percent of the parent material strength. It is fast, clean and it seals.

The fixed cost is the barrier. A welder for a small part costs EUR 15,000 to EUR 30,000, plus a horn and a fixture for each part, plus the tooling change to add the energy director and the joint design. That is why ultrasonic welding appears at high volume and rarely below it.

The design requirements are specific. The joint needs a uniform energy director, a register to align the two halves, and a way for the melted material to flow without flashing out. A part designed for screws and then welded will flash, misalign and produce inconsistent joints, and the fix is a tool change rather than a process setting.

The material matters as well. Amorphous plastics such as ABS and polycarbonate weld well; semi-crystalline plastics such as PP and PE absorb the vibration and are difficult. Glass filled materials are harder again, because the filler does not melt and interferes with the bond.

Adhesive: seals and needs preparation

An adhesive is the most flexible of the four in terms of what it can join. It can bond dissimilar materials, it can fill gaps, it can seal, and it needs no feature in the moulding beyond a flat surface. It is also the only method that can join a plastic to a metal without an insert.

The cost is in the process. The surface has to be clean and often treated, the adhesive has to be dispensed in a controlled bead, and the parts have to be held while it cures. A cure that takes 24 hours at room temperature needs a warehouse full of fixtures; a cure that takes 30 seconds under UV needs the joint to be transparent to UV, which most assemblies are not.

Adhesive is also the method most sensitive to operator variation. A bead that is 20 percent too thin leaves a gap that does not seal; a bead that is too thick squeezes out and makes a mess. Automated dispensing solves it and adds the fixed cost that puts adhesive in the same bracket as ultrasonic welding.

Where adhesive wins is on mixed materials and on parts that cannot be redesigned. A plastic lens bonded into a metal bezel, a silicone seal bonded to a housing, a display bonded to a front panel: these are adhesive jobs, and no amount of snap fit design will replace them.

Choosing, and what disassembly costs

The decision tree has three questions. Does the joint have to come apart? If yes, screws or snap fits, and screws if it has to come apart more than twice. Does the joint have to seal? If yes, ultrasonic welding or adhesive, and adhesive if the materials differ. Is the volume above about 50,000 a year? If yes, the fixed cost of welding or dispensing can be absorbed.

Disassembly deserves more attention than it usually gets. A product that cannot be opened cannot be repaired, and increasingly it cannot be recycled either. A welded housing has to be shredded as a unit; a screwed housing can be separated into its materials in seconds. That is a design decision with a cost that appears years later.

Table 1 — The four methods side by side on a small housing

Table 2 — Starting from the requirement rather than the method

The two tables are meant to be read together: the first says what each method costs, the second says which one the requirement points at. The table is a starting point, not an answer. The right method for a given part depends on the wall thickness available, the material, the service requirement and the volume, and the only reliable way to choose is to cost all four against the actual geometry before the tool is cut.