Embedding Electronics in 3D Printed Parts: Creepage, Inserts, Potting and Heat That Stay Safe

Every functional print starts the same way: you model a part, you print it, you hold it, and then you realise it has to talk to something. A switch, a motor, a sensor, a battery, an LED strip. At that moment the mechanical design stops being the whole story. Printed plastic is an electrical environment too, and it is an environment with rules that almost nobody explains before the first part fails.

InnoStation 3D supplies ANTINSKY materials and publishing notes on this topic usually stop at which filament conducts and which does not. That is the least interesting question. The questions that decide whether your housing survives a year are about geometry: where the conductor is allowed to go, how much clearance the voltage demands, what heat does to a cavity sealed inside plastic, and how a wall that is a perfect insulator at 5 volts becomes a liability at 230 volts.

Why Printed Plastic Is an Electrical Material First

Thermoplastics are dielectrics. PLA, PETG, ABS, ASA and TPU all resist current flow, and their bulk resistivity is high enough that a multi-layer wall is effectively an open circuit at hobby voltages. That is the good news, and it is where most advice ends.

The bad news is that dielectric strength is not a single number you can read off a filament spec sheet, because a printed wall is not a solid material. It is a stack of extruded roads with air gaps, Z-seams, and unbonded interfaces between them. What matters electrically is the weakest link in that stack, not the average of the good ones.

• Layer interfaces are the thinnest electrical barrier in any FDM part. Where two layers meet without full diffusion, you get a partial air gap. Air breaks down at roughly 3 kV per millimetre, which sounds generous, but the gap is often only microns tall and the local field can concentrate at the seam edge.
• Internal cavities left by infill are electrical voids. A 15 percent gyroid infill is mostly air with a tortuous path of plastic through it. That is fine until you put a high-voltage trace on one wall and a grounded fastener on the other.
• Moisture absorption quietly raises a wall's conductivity over time. Most engineering filaments absorb between 0.2 and 2 percent of their weight in water depending on polymer, and a wet wall is measurably less insulating than a dry one.
• Carbon and metal fillers change the picture completely. Carbon-fibre and carbon-nanotube loaded filaments are engineered to be somewhat conductive; using one as a structural housing around bare conductors is a design error, not a shortcut.

The practical rule is that printed plastic is an excellent insulator at low voltage and a questionable one anywhere near mains potential. Design as though the wall is a barrier with a weak seam in it, because it is.

The Voltage Clearance Table

Creepage is the distance current must travel along a surface. Clearance is the distance it must jump through air. Printed parts are hostile to both, because layer lines create a continuous rough staircase along any surface, and roughness effectively shortens the path a contaminant film can bridge.

These figures are conservative design minimums for enclosed, clean, dry printed assemblies. They are a starting point for hobby and light commercial work, not a substitute for the electrical safety standard that governs your product category.

Working voltage	Minimum wall between conductors	Minimum air gap	Reality check
Up to 12 V DC	0.8 mm, one or two perimeters	0.5 mm	Almost any print passes. Focus on mechanical fit instead.
24 V DC	1.2 mm	0.8 mm	Arcing is not the risk. Damaged insulation chafing on a rough wall is.
48 V DC	1.5 mm	1.0 mm	Above 30 V, treat every exposed conductor as a shock hazard.
120 V AC	2.5 mm	1.5 mm	Layer seams and moisture matter. Do not rely on a two-perimeter wall.
230 V AC	3.0 mm minimum, 5 mm preferred	2.0 mm	Voids, seams and contamination all reduce effective spacing. Verify, do not assume.
Above 230 V	Do not design it in printed plastic	Use a certified enclosure	The material choice, not the geometry, is the limiting factor.

Two design consequences follow immediately. First, wall thickness for electrical separation should always be counted in whole perimeters, never in slicer fill percentage, because infill voids are not a barrier. Second, that dimensional error you tuned out for mechanical fit reappears on the electrical side: a 2.5 mm specified wall that prints at 2.3 mm is a 10 percent reduction in creepage distance.

Three Way

Conductors Along a Printed Channel

Embedding a wire inside a wall is the most seductive idea and the most commonly botched. There are three workable methods, and they fail for different reasons.

Method	How it works	Failure mode	Best for
Printed channel and press-in	Model an open channel, print, press the wire in, cap it	Wire chafes on the rough channel floor; channel collapses on overhangs	Rigid single-core wire in low-flex joints
Pause and embed	Print a floor, pause, lay the wire, resume over it	Nozzle hits the wire; wire melts if the layer above is hot	Flat runs in one plane, no crossings
Printed into an open groove	Half-open groove on an outer face, closed with a cover or a bead of sealant	Cosmetically visible, incidental damage in handling	Serviceable rework and prototypes

Whichever you choose, size the channel for the wire plus insulation plus a clearance allowance, not the bare conductor. A 1.5 mm outside diameter wire wants a 1.7 to 1.8 mm channel so it slides without being forced, and the channel floor should be printed with a solid bottom layer stack, not bridging over infill. Where a wire must turn a corner inside a channel, give the turn a generous radius; sharp printed corners become wire-cutters over many insertion cycles.

Strain relief belongs at every point the wire leaves the plastic. A printed cable entry should have a graduated grommet, either a printed one or a commercial rubber insert, and the geometry should clamp the insulation rather than the conductor.

Current-Carrying Inserts

Threaded inserts, rivets, and press-fit pins are all electrical penetrations. A metal insert passing through a wall is a conductor crossing your barrier, and it usually carries whatever potential the fastener picks up from the mating part.

Two rules keep this manageable. Insulate the insert from any conductor inside the housing by at least the creepage distance in the table above, measured from the closest point of the insert to the closest point of the live conductor, following the surface. And remember that the inserts you sized for mechanical pull-out strength also need a temperature allowance: heat-set inserts require the surrounding plastic to reach 200 to 250 degrees C locally, which is far above the glass transition of PLA and can distort thin electrical walls near the boss.

Printed Conductive Traces

Conductive filaments exist, and they are consistently misunderstood. They are composites of a polymer with carbon black, carbon fibre, graphene, or metal powder. Their resistivity is orders of magnitude higher than copper, in the range of a few ohm-centimetres rather than micro-ohm-centimetres, and it varies with print orientation and flow.

That makes them useful for electrostatic dissipation, shielding, touch sensing, and heat, and useless for power distribution. A printed trace the same cross-section as an 18 AWG wire will have hundreds of times its resistance and will produce a voltage drop and a hot spot you did not budget for. If you need to carry current, carry it in metal. Use printed conductive paths only where the required current is milliamps or the required function is field control rather than energy transfer.

Insulation, Potting and Sealing

Sealing a cavity is where electrical design meets the watertight problem, and the two have opposite intuitions. Watertight printing is about eliminating capillary paths through layer lines, which is a mechanical discipline of wall count, temperature, and flow. Electrical encapsulation is about eliminating voids around conductors, which is a materials discipline.

Potting a printed cavity with a two-part epoxy or polyurethane serves four purposes at once: it excludes moisture, it immobilises conductors so they cannot chafe, it raises the dielectric strength of the void, and it conducts heat away from hot components. The catch is that most potting compounds cure exothermically, and a large pour in a thin printed shell can reach temperatures that soften the shell before it sets.

1. Print the cavity with at least three solid perimeters and three solid top and bottom layers. The shell must hold the pour and resist its own thermal load.
2. Leave a fill port and a vent port, at opposite corners. A single port traps air and produces a void exactly where you wanted encapsulation.
3. Pour in lifts rather than one shot. Two or three pours with a cure interval between them keeps the exotherm manageable.
4. Confirm the potting compound's dielectric strength and its service temperature overlap your worst case, not your typical case.
5. Pot only what you will never service. Once encapsulated, the assembly is a single non-repairable unit.

Where full potting is overkill, a conformal coating on the printed surface near conductors is a cheap middle path. It seals the surface pores, reduces moisture uptake, and improves surface resistivity without filling the cavity.

Heat Is the Real Constraint

Every electronic component in a printed housing is a heat source in a thermal insulator. That is the fundamental conflict of embedded electronics, and it does not appear in any filament comparison table.

TPU is the worst case, with a practical service window that starts softening far below the temperatures a small motor driver or a linear regulator will happily reach. PLA is little better. PETG and ABS hold up longer, and the glass transition temperature, not the melting point, is the number that governs when a housing starts to sag under load. A wall that creeps under sustained heat is exactly the long-term deformation failure mode that shows up months after the build, and electronics housings are where it hides best because nobody reopens a sealed box to look.

The mitigation is geometric before it is material. Give the heat source a conduction path to the outside wall, ideally a solid printed boss or a metal insert that reaches the exterior. Add vent slots where the component is not moisture-sensitive, and place them high and low to encourage a chimney effect. Keep thermally sensitive plastics away from the hottest component, and if the cavity must be sealed, either accept a lower duty cycle or move to a material with a genuinely higher glass transition temperature.

Component	Typical dissipation	Practical housing requirement
LED indicator	Under 0.1 W	Nothing. Any filament is fine.
Microcontroller board at idle	0.2 to 0.5 W	A small vent or a conductive path to the wall.
Linear voltage regulator	0.5 to 2 W	Metal tab bonded to an exterior wall or insert, plus vents.
Small DC motor under load	2 to 10 W	Open structure or a metal heat path. Sealed PLA will soften.
Li-ion cell under charge	Self-heating plus fault energy	Vented cavity, no sealed encapsulation, flame-retardant material.

Material Choice for Electrical Housings

Filament selection for an electrical enclosure is driven by three properties in order: thermal service limit, flame behaviour, and dimensional stability. Electrical resistivity is a distant fourth, because most unfilled thermoplastics are already adequate insulators at low voltage.

• PETG is the pragmatic default for low-voltage enclosures. It prints easily, resists moisture better than PLA, and tolerates the warmth of a lightly loaded board. It is not flame retardant and it will creep under sustained load and warmth, so keep it away from mains and from anything that gets genuinely hot.
• ABS and ASA offer a higher service temperature and better long-term dimensional stability, at the cost of requiring an enclosure to print. Where a housing will see continuous warmth, this is the honest upgrade over PETG. There is a real trade-off between open and enclosed printing that decides whether ABS is even available to you.
• Flame-retardant grades exist in both filament and resin form and are the correct choice for anything with a lithium cell, a mains-derived rail, or a fault current. A flame-retardant filament such as ANTINSKY ABS FR filament adds halogen-free flame inhibition to the ABS thermal profile, which is exactly the combination an enclosure wants. In resin, a flame-retardant black resin covers the same ground for small precision housings.
• Anti-static and ESD-safe materials solve the opposite problem: bleeding charge away instead of blocking it. An ESD-safe anti-static resin belongs around boards that handle static-sensitive components, and an ESD-capable filament such as Polymaker Fiberon PA612 ESD is the filament equivalent. Note the direction of the requirement, because it inverts everything else in this article: these materials are deliberately slightly conductive, so they must never be your primary insulation.
• Impact and toughness grades are for enclosures that get dropped. A black engineering impact resin or a high-tough resin gives a housing that survives handling far better than a brittle standard resin, and impact resistance matters more than dielectric strength for a part that lives in a bag or a toolbox.
• Standard structural filaments remain the right answer for the bulk of load-bearing enclosure work where no thermal or flame requirement applies. A PETG filament for general housings and an ABS filament where temperature matters cover most of what a workshop actually builds, and TPU is the pick for grommets, strain reliefs and shock mounts where flexibility is a functional requirement rather than a preference.

One material caveat that deserves its own line: never use a carbon-fibre or metal-filled filament as the insulating wall around a live conductor. These materials are engineered for stiffness and, in some cases, for conductivity. Using one as an electrical barrier is a category error that will not announce itself until conditions change.

Design Rules Checklist

1. Separate mechanical walls from electrical barriers. An electrical barrier is solid plastic, in whole perimeters, with no infill voids in the path.
2. Size creepage and clearance from the voltage table, then add margin for print tolerance and moisture.
3. Never route a wire through infill. Channels need solid floors and ceilings.
4. Give every entry point a grommet or a printed strain relief that grips insulation, not conductor.
5. Keep fasteners and inserts clear of live conductors by the full creepage distance.
6. Provide a real thermal path from any dissipating component to the exterior wall.
7. Keep lithium cells in a vented cavity, never in a potted one.
8. Choose flame-retardant material wherever a fault current is possible.
9. Use ESD materials around sensitive boards, and only there.
10. Prototype the electrical assembly in the printed part before committing the final housing geometry. Clearance problems are cheap to fix in a draft and expensive to fix in a moulded or potted unit.

Where the Mechanical and Electrical Designs Meet

Embedded electronics in printed parts is not a separate discipline bolted onto mechanical design. It is the same discipline under a second set of constraints. The wall you thickened for stiffness now has to be solid rather than infilled. The press fit you tuned with the snap and press fit guidance now has to keep a conductor at a fixed distance as well as a fixed load. The adhesive joint you chose by joint geometry and adhesive chemistry now also has to survive the warming and cooling cycle of the electronics inside it.

Two areas repay attention specifically. If the part must keep water out as well as current in, the capillary leak problem is your real opponent, because a channel that carries a wire also carries water along that wire. And if the housing uses multiple materials for mechanical or cosmetic reasons, the purge and transition behaviour determines whether the interface between two polymers is a continuous barrier or a seam waiting to open.

The short version: design the electrical path first, because it is the constraint that cannot be negotiated later. Geometry, material and enclosure are all negotiable. Creepage distance and thermal headroom are not.