Choosing how to seal a connector is a manufacturing-economics decision before it is a materials one — driven by annual volume, tooling, and depth:
Key Takeaways
- The choice is set by annual volume and tooling budget, not by which method is "best" — potting, overmolding, and LPM each win in a different volume band.
- Potting wins below ~500 units/year and for deep-submersion pressure: near-zero tooling and rigid epoxy that resists hydrostatic deformation, at a high per-unit labor cost.
- Overmolding wins above ~5,000 units/year: expensive steel tooling amortizes into the lowest unit cost, and strain relief is molded directly into the shape.
- Low-Pressure Molding (LPM) bridges the mid-volume gap (1k–10k/year) with cheaper aluminum tooling and pressures low enough to seal fragile PCBs and solder joints.
- Every method fails at the same place — the cable-to-seal interface — so the seal is only watertight when the encapsulant chemically bonds or mechanically locks to the jacket.
Engineering rule of thumb: pick the process by annual volume first — pot below 500, low-pressure mold through the mid-thousands, overmold above 5,000 — then choose the material and rating within that process.
Encapsulation Is a Volume and Tooling Decision
Before debating resins or IP ratings, the encapsulation method for a sealed custom cable assembly and wire harness is decided by two numbers: the tooling budget and the annual build volume. Potting carries almost no tooling cost but a high per-unit labor cost; overmolding inverts that, with a steep steel-mold investment that only pays back across thousands of units.
The crossover is the whole decision. At low volume, there is no production run to amortize a steel mold against, so potting's labor cost is cheaper overall. As volume climbs, the molded unit cost falls until the amortized tooling is trivial per part. Low-pressure molding sits deliberately in between, using inexpensive aluminum tooling to capture much of overmolding speed at a fraction of its setup cost.
Potting: Low Volume, High Pressure and Depth
Potting is a chemistry-first process: a potting cup or backshell is filled with a two-part resin that cures into a solid encapsulant. It needs no injection mold, which makes it the default for prototypes and runs under a few hundred units, and its rigidity makes it the strongest choice for pressure.
- Epoxy: Hard, rigid, chemically resistant — the choice for deep-submersion and high-pressure applications, because a rigid fill resists the hydrostatic deformation that can compromise a softer seal. The risk is thermal-shock cracking if the resin's CTE mismatches the housing.
- Urethane: Flexible, tolerant of vibration and thermal cycling, and bonds well to many housing plastics.
- Silicone: High-temperature capability but notoriously poor adhesion — it bonds to almost nothing without expensive primers.
The dominant failure mode is trapped air. Hand potting almost always leaves microscopic voids around the conductors, and under submersion those voids channel water. A true IP68 submersion-rated waterproof cable assembly therefore requires vacuum (degassed) potting, not open-air pour.
Overmolding: High Volume, Molded Strain Relief
Overmolding is a pressure-first process: molten polymer is injected at high pressure so it forces into every crevice of the connector back-end and cures in seconds. The steel tooling is costly, but at volume the per-unit cost is the lowest of the three methods, and the process delivers something potting cannot — a strain relief molded directly into the part.
Because a potted block transitions abruptly from rigid fill to bare cable, it creates a stress-concentration point where the wire exits; an overmold can instead be shaped into a graduated flex tail. The resin choice that makes or breaks the seal — TPU versus PVC versus Santoprene, and matching that chemistry to the cable jacket — is a separate decision covered in our overmold material selection guide. The principle to carry into this process decision is that high injection pressure can crush a connector or solder joint not rated for it.
Low-Pressure Molding: The Mid-Volume Middle Ground
Low-pressure molding injects a hot-melt polyamide adhesive at a fraction of standard overmolding pressure — low enough that it will not crush sensitive components. It cures in seconds like overmolding but uses inexpensive aluminum tooling closer to potting's cost, which is why it owns the mid-volume band.
- Best use: Sealing exposed PCBs, fragile solder joints, or delicate sub-assemblies that standard high-pressure overmolding would damage.
- Economics: Aluminum tooling and second-scale cycle times make it viable from roughly 1,000 to 10,000 units a year — exactly the range where potting is too slow and steel-tool overmolding has not yet paid back.
Matching the Process to the Sealing Requirement
Once the process is chosen on volume and pressure, confirm it against the sealing target. The encapsulation method is only one of several ways to reach a given ingress rating, and which method (including heat-shrink) best reaches washdown-grade sealing is compared in our IP69K cable sealing guide.
It is also worth fixing the target rating precisely before committing tooling, since IP67, IP68, and IP69K each certify a different exposure — submersion depth versus high-pressure high-temperature washdown — as detailed in our guide to IP67 vs. IP68 vs. IP69K ratings. Specifying a deep-submersion rating points toward rigid potting; a washdown rating with high cycle volume points toward overmolding.
Encapsulation Process Comparison
| Feature | Potting (Encapsulation) | Overmolding (Injection) | Low-Pressure Molding |
|---|---|---|---|
| Tooling Cost | None / low | Highest (steel molds) | Moderate (aluminum molds) |
| Unit Cost | High (labor-intensive) | Lowest (automated) | Moderate |
| Volume Sweet Spot | Low (< 1k/yr) | High (> 5k/yr) | Mid (1k–10k/yr) |
| Strain Relief | Poor (rigid transition) | Excellent (molded) | Good |
| Depth / Pressure | Excellent (rigid fill) | Good (material dependent) | Fair |
| Cure / Cycle Time | Hours (24h+) | Seconds | Seconds |
| Primary Risk | Air voids / mix ratio | Delamination / component crush | Material softness |
Frequently Asked Questions (FAQ)
Can I overmold a standard off-the-shelf connector?
Not easily. Standard connectors such as a generic USB or RJ45 are not built for injection-molding pressure — the plastic can crush internal contacts or flash into the mating face. You generally need "overmold-ready" connectors with sealing ribs and dams to block plastic ingress, or you step down to low-pressure molding, which will not damage the connector.
Is potting waterproof without a vacuum?
Not reliably. Open-air "hand potting" may repel splashing (around IP65), but it almost always traps microscopic air bubbles around the wires. Under submersion, those voids channel water to the contacts. For a true IP68 submersion seal, vacuum (degassed) potting is mandatory.
What happens if the encapsulant does not bond to the cable?
This is the most common failure in sealed assemblies. If you overmold or pot onto a jacket it cannot bond to — PTFE or silicone, for example — water wicks by capillary action along the jacket-to-seal interface straight to the contacts. The fix is either matching the materials for a chemical bond or designing mechanical interlocks (grooves and holes) the encapsulant flows around.
Which process is best for repairability?
None of them. Potting and overmolding are both permanent — cured epoxy or molded TPU cannot be removed without destroying the connector. If field repair or rework is a requirement, use a mechanical backshell with a compression gland or grommet instead of an encapsulation process.
Sealing a connector is a process decision driven by volume and pressure before it is a material decision. Pot at low volume or deep submersion; low-pressure mold through the mid-volume range and over fragile components; overmold at high volume where molded strain relief and the lowest unit cost matter most. Fix the build volume and the ingress rating first, and the encapsulation process — and the material and tooling that follow — falls out of those two constraints.