When the Part Has to Be Soft and Tooling Is Far Off

A gasket that keeps water out, a grip that is pleasant to hold, a pad that touches skin, a flexible button membrane, a strain-relief sleeve on a cable — most products have at least one part that cannot be printed in rigid plastic and cannot be skipped during testing. Silicone casting is the fast way to get exactly those parts in small quantities, without cutting a production mold and without compromising on how the material feels.

One distinction to get straight immediately: here the silicone is the part. That is different from silicone-mold casting, where a silicone tool is used to reproduce rigid parts — the method described in urethane casting for low-volume production. Same material, opposite role.

How the Process Actually Works

  1. Master model. Build a model of the desired part, usually resin printed at high resolution and sanded smooth. Every surface flaw is copied into every casting, so master quality decides part quality — this is one of the places where choosing between FDM and SLA genuinely matters.
  2. Mold construction. A two-part mold is built around the master, typically in rigid resin, machined aluminum for longer runs, or a harder silicone, with a sprue and vents at the high points.
  3. Mix and pour. Two-part silicone is weighed precisely, degassed under vacuum to pull out entrained air, and poured slowly into the mold. A wrong ratio or incomplete mixing gives you a part that never fully cures — tacky, weak, and unusable.
  4. Cure and demold. After the cure time, the part is removed, flash is trimmed, and dimensions are checked against the drawing. Heat post-curing speeds things up and improves properties on many systems.

One mold typically yields dozens of pulls before it degrades — more than enough for a test series, demo units and a first customer round. End to end, from a finished CAD file to parts in hand, the process is measured in days rather than weeks, which is exactly what makes it so useful mid-development: you can discover on Monday that the seal leaks, fix the model, and hold a corrected version by Thursday.

Choosing the Material: Start With Durometer

The first parameter to fix is hardness on the Shore A scale, and it drives everything else:

  • Very soft, roughly Shore A 10 to 20 — skin-contact pads, cushioning, tissue-simulating models.
  • Mid range, roughly Shore A 30 to 50 — gaskets, straps, grips, flexible buttons. Most prototype soft parts live here.
  • Firmer, roughly Shore A 60 to 80 — wheels, feet, bumpers, vibration-damping mounts.

The full scale and how to interpret supplier data sheets is covered in our Shore hardness guide. The best feature of this process is that you can cast the same geometry in three different durometers and let users choose — something essentially impossible once a production tool exists.

Other parameters to specify: platinum-cure versus tin-cure chemistry (platinum systems are more accurate and stable but are inhibited by sulfur-containing clays, some 3D printing resins and latex, which causes maddening surface cure failures), clarity or pigment, service temperature range, tear strength and elongation, and — for anything with prolonged skin contact — a material rated for that use. For medical applications, biocompatibility testing is a formal process, not a supplier claim; see ISO 10993 biocompatibility testing. The wider material selection framework is in how to choose materials for a new product.

What to Design In Advance

  • Release direction. Silicone is flexible and forgiving, so moderate undercuts can be pulled without a side action — a real advantage at prototype stage that disappears in production tooling.
  • Wall thickness. Too thin and the cavity will not fill; too thick and the cure is slow and prone to trapped bubbles. Keep sections reasonably even.
  • Air bubbles. The number one defect. Vacuum degassing before pouring, pouring in a thin stream, and putting vents at every high point solve most cases; pressure casting solves the rest.
  • Bonding. Silicone does not stick to ordinary adhesives. If the soft part must attach to a rigid one, design mechanical capture — a groove, a barb, a lip, a stretched cuff — or plan for a primer system. Assuming glue will work is a very common and very costly late surprise.
  • Shrinkage. Small but real. On critical dimensions, measure the first casting and correct the master.
  • Sealing geometry. If the point of the part is keeping water out, the gland design matters more than the material — the wider approach is in our article on building a waterproof prototype.

Alternatives Worth Comparing

For a single quick check, printing a flexible part in TPU can be faster and cheaper than building a mold, at the cost of surface finish, sealing quality and true elastic behavior — see TPU 3D printing for flexible parts. Die-cut sheet gasket material is even faster for simple flat seals. Cast silicone earns its place when the geometry is three-dimensional, the feel matters, or you need multiple identical units.

When to Stop and Move to Real Production

Hand casting is excellent for tens of units and does not scale much beyond that. Once you need hundreds or thousands, consistency between parts becomes the limiting factor and labor cost dominates: liquid silicone rubber injection molding or compression molding becomes the right answer, with real tooling cost and much lower per-part cost. If the soft part is bonded to a rigid one in production, the target process is usually overmolding or two-shot molding, and it is worth designing the prototype so the geometry can migrate there without a redesign.

Get the Soft Parts Made Properly

Soft components are where prototypes leak, rattle and feel cheap, and they are usually the last thing anyone plans for. If your product needs seals, grips or pads in small quantities, send us the details through our contact form and we will specify the durometer, the mold approach and the number of pulls you actually need.