Choosing Materials for a Hardware Product

The strain sum below is hypothetical teaching maths. The allowable strains are assumed for the exercise. They are not a grade datasheet, not a BrahmWorks test and not a permissible limit you can copy onto a drawing. The grade supplier's data, read for the temperature and the chemical you actually have, replaces them.

Split the duty into questions a material can fail. What load, how often, and is it a one-time assembly or a daily opening? What temperature, continuous and peak, including the inside of a closed box and a delivery truck? What chemicals: skin, sweat, disinfectant, oil, battery leakage? Sunlight or only indoor light? A flame or a contact-safety requirement, which is a grade and a product question, not a family slogan? Appearance: gloss, texture, colour stability, scratch? Will someone bond, paint, weld or print on it?

Add the process in the same note. Mould, machine, sheet, extrude, or print-as-production. The process deletes families faster than the brochure does. A living hinge points at polyolefins. A clear lens points at a short list, with scratch and chemical resistance still open. A structural bracket you will machine in tens may never need a moulding grade at all.

Food contact, medical contact, children's products and flame requirements are a named-grade review with a qualified person, not a line in a family table. This article clears none of those uses.

Use this as a way to ask, not as a ranking.

Polypropylene is often where a living hinge and broad chemical resistance are doing the work. It is less stiff than many housings want, and it is a poor candidate when you need to bond or paint without a specific treatment. ABS is a common housing family: sane cosmetics, easy to mould, weaker in sunlight and in some chemicals than people expect from a "general" plastic. Polycarbonate is chosen for impact and for heat, and it is known to be sensitive to environmental stress cracking in contact with some solvents and oils. That is a reason to name the chemical, not a reason to ban the family. PC/ABS blends are a compromise on cosmetics, heat and moulding. The blend ratio matters, so the family name is not the specification.

Polyamide, nylon, is tough and useful in wear and in clips, and it takes up moisture. Dimensions and stiffness move between the dry part you measured in winter and the conditioned part in a humid room. If the stack-up cannot stand that move, do not discover it after the tool. A glass-filled grade stiffens and shrinks differently from the unfilled one, and it wears the tool and the mating surface. Elastomers and thermoplastic elastomers are for seals and grips. The question is compression set and the chemical, not the softness in the hand sample. Aluminium moves heat and gives stiffness in thin walls; it also changes a radio, needs a finish, and has its own corrosion questions. Sheet steel is often the cost and EMI answer, with edges, paint and fasteners as the real design.

None of these sentences is a measured property. A grade inside the family can break the stereotype. Specify the grade, the colour package and the filler, or you have not specified a material.

Record it as a stand-in. A printed clip does not qualify an ABS clip. A machined acetal latch does not qualify a polypropylene hinge. You may use the stand-in to learn geometry, kinematics and assembly. You may not transfer a pass on strength, creep, chemical resistance or a snap that depends on strain. When you change material, list the tests that are void. Keeping the old report in the pack without that list is how a later review thinks the duty was met.

Write the supply rule on the same page. One colour concentrate from one plant is a different risk from a grade with a second source. A boring grade you can buy again beats a perfect grade you cannot reorder. Ask the lead time. Do not invent one.

Invented product: a wall device with a polymer door over a battery, wiped in service with a disinfectant the customer already uses. Two different duties, often forced into one resin because it is convenient.

The housing wants cosmetics and a mouldable wall. The door wants many openings and a chemical splash. One resin can do both only if that grade actually survives the wipe and the strain. The teaching split is: pick the housing family for moulding and appearance, and check the disinfectant against that grade before you fall in love with the texture. Design the door as a replaceable part if the strain sum says the housing resin cannot take repeated openings. Do not "add a fillet later" as a substitute for that check. No wipe test and no cycle test are reported in this article.

For a straight cantilever of constant rectangular section, in elastic bending, small deflection, the outer-fibre strain at the root is:

Strain = (3 × thickness × deflection) / (2 × length²)

The length is the flexible length, not the overall finger including the hook. Assume, for the exercise only, thickness 1.2 mm, length 15 mm, deflection at the hook 1.5 mm.

Numerator = 3 × 1.2 × 1.5 = 5.4. Denominator = 2 × 225 = 450. Strain = 5.4 / 450 = 0.012, which is 1.2 percent.

Now assume, again only for teaching, that a one-time assembly may use up to 3 percent strain and a latch opened often may use only 0.8 percent. Those ceilings are not data. Under them, 1.2 percent is inside the one-time ceiling and outside the repeated-use ceiling. A door opened in the field fails this assumed rule. A clip that is assembled once in the factory and never touched passes it. The geometry can stay if the duty is one-time, or the finger must get longer or thinner, or the material must be one whose real allowable strain, from the grade, covers 1.2 percent at the temperature and the chemical.

Check the direction of the change. Length is squared. Taking length from 15 mm to 18 mm, same thickness and deflection: denominator = 2 × 324 = 648. Strain = 5.4 / 648 = 0.0083, about 0.83 percent, still on the wrong side of an 0.8 percent ceiling. Length 19 mm: denominator = 2 × 361 = 722. Strain = 5.4 / 722 = 0.0075, about 0.75 percent, which would sit under that assumed ceiling. A millimetre on a short finger is not a detail. It is the design. The formula also ignores stress concentration at a sharp root and the hook's own stiffness. A sharp corner makes the real strain worse than this number.

When the material finally chosen voids the prototype evidence, the production path has to say so. That path is How to Take a Hardware Prototype to Production.

• Load, cycles, temperature, chemical, light and appearance are written.
• The manufacturing process is named beside the material.
• Flame, food and body-contact claims are either out of scope or on a named grade review.
• The family is a reason. The specification is a grade, filler and colour.
• Moisture, chemical stress cracking and creep are asked, even if the answer is "not this duty".
• The prototype material is labelled as a stand-in, with the tests it cannot support.
• A second source, or the decision to live without one, is explicit.
• Paint, adhesive and printing are checked against the grade, not assumed.
• Snap or clip strain is estimated, then replaced by grade data.
• A change of grade voids the listed tests until they are repeated.
• Cost is compared after the duty filter, at the real quantity.

You can pick the grade late inside a family if shrink and chemical behaviour stay compatible. You cannot treat "plastic" as a placeholder and design the wall, the draft and the snap as if they were universal. The tool is cut for a shrink. The snap is shaped for a strain. Both are material decisions. Leave the grade open only where you have written the band of shrink and stiffness you are still willing to accept.

No. Metal solves stiffness and heat and often shielding. It brings finish, corrosion, edges, weight, and a radio that now sits in a can. A polymer housing that meets the duty is not a compromise. A metal housing that was never designed for the antenna is not an upgrade. Compare on the duty, including assembly and the coating.

It can be a legitimate specification. It changes melt flow, colour and sometimes impact, and it varies by source. If you allow it, cap it and inspect the properties you care about. If you need a cosmetic face or a safety-relevant impact behaviour, "regrind allowed" with no cap is an open material change on every batch. This is not an environmental claim. It is process control.

Put the grade on the drawing, state whether equivalents are forbidden, and require a deviation when they are not. "Or equivalent" without a test for the duty is permission. A lookalike that matches a datasheet headline and fails the disinfectant, or shrinks differently, is a new design. Approve it as one, or refuse it.

Bring the environment note, the process, the grade you think you want, and the list of tests run on a different polymer. BrahmWorks can say which of those tests still stand, and which choice is still open.

Review the next stage with BrahmWorks