IEC 60664-1: Where Your Clearance Numbers Come From

Two suppliers quote the same current, the same copper, the same enclosure, and their drawings show 12 mm in one case and 18 mm in the other. Neither is arbitrary. IEC 60664-1 derives both figures from a short list of declared inputs, and the two suppliers simply declared different ones.

The standard is the insulation coordination reference that IEC 61439-1 invokes. It does not describe busbars; it describes how to arrive at distances that will not break down at the impulse voltages and the continuous voltages a low-voltage assembly actually sees. That makes it the single most useful standard for anyone arguing about a gap.

The practical consequence: a clearance number without its inputs is not a specification, it is a habit. Ask for the four inputs and you can reproduce the number yourself, in either direction.

If you have not yet settled the assembly-level question, start with the standards hub; the verification documents that consume these distances are covered in the IEC 61439-1 guide.

Everything on this page follows from these declarations. Get them wrong and the distances are wrong even if the arithmetic is right.

• Rated insulation voltage U_i — the voltage the assembly is designed to withstand continuously. Not the system voltage, and not the rated operational voltage, though in practice they are often equal.
• Overvoltage category — I to IV, describing the transient overvoltage the circuit is expected to survive. A feed direct from the utility service entrance sits higher than a circuit behind a surge arrester and a transformer; the category is what converts that into a test impulse level.
• Pollution degree — 1 to 4. In an enclosure that is cleaned and closed, 2 is the normal assumption. Condensation, conductive dust, or salt mist pushes it up, and the distances grow accordingly.
• Material group — derived from the comparative tracking index of the insulating surface. Group I is the most resistant, IIIb the least. This input only affects creepage, not clearance.

The correction is altitude. Air thins with height, so clearance must grow above 2,000 m. Below that, the factor is unity and most projects never touch it. Above it — and plenty of Andean, Ethiopian and Chinese plateau projects sit above it — the clearance table is no longer valid as printed.

Write all five onto the drawing, next to the distances they produced. It costs one line and ends most arguments before they start.

Clearance protects against a short, steep event — a switching or lightning impulse arriving through the system. So the table is indexed by the impulse withstand voltage required for the declared overvoltage category at the declared U_i, and by pollution degree. Working voltage does not enter it directly.

This is the single most common misunderstanding in busbar drawings. A 400 V system at overvoltage category III needs a 4 kV impulse level for 300 V nominal, which lands near 3 mm minimum in air at pollution degree 2 — yet 400 V working voltage suggests a much larger creepage. The two numbers serve different events and must be looked up separately.

Distance applies to the shortest path through air between conductors at different potential. That includes paths that wander: past a bolt head, around a standoff, between a conductor edge and the wall of an enclosure. Measure the shortest one, not the one that looks deliberate.

Also check the path that is not on the drawing: a busbar with a 3 mm edge-to-bolt-head gap can pass the drawing review and fail the first article because the washer sits 0.5 mm differently.

Creepage protects against a slow event — leakage tracking across a wet, contaminated surface. It is indexed by working voltage (RMS, long term), pollution degree, and material group. This is where a supplier's choice of film or coating changes a number the customer never sees.

Three mechanics matter when you measure it:

• Grooves and ribs count only under conditions. A narrow groove is treated as a straight line across it, which helps; a wide groove follows its contour, which does not. Ribs help more than grooves and the tables assume a specific minimum height.
• Joints interrupt the path. Two parts pressed together without cement form a joint that may be counted as a barrier — but only if the standard's conditions on width and contact are met, and only if nothing shifts in service.
• The worst surface controls. If a laminate carries an external tracking-resistant layer and an exposed cut edge with a much lower index, the cut edge governs the declared material group unless the layer is shown to cover it.

For a laminated stack the interesting surfaces are usually the ends: cut laminate edges, the film edge, potting compound, and any conformal coating that the drawing shows but the process does not reliably produce.

The tables assume a surface in air. A laminated busbar is mostly a surface in solid dielectric, and that changes three things.

1. The layer-to-layer gap is not an air gap. Between copper layers sits PET, Nomex or a DLF film, bonded under heat and pressure. Distances inside the stack are governed by solid insulation rules and by partial discharge behaviour, not by creepage tables. Quoting an air clearance for an internal layer gap is a category error — see partial discharge in laminated stacks.
2. The exposed edge is the weak surface. A cut laminate edge can expose film plies, adhesive and copper burr. If the declared material group rests on the film's tracking index, the edge needs coverage — an insulating cap, potting, or a controlled edge distance to the nearest different potential.
3. Hardware changes the shortest path. Bolts, washers, standoffs and cable lugs sit in the space the drawing measured as air. The distance that matters at commissioning includes them, and the distance the assembler can actually build includes tolerances.

None of these argues for a bigger number everywhere. They argue for stating which insulation system covers which gap, so a reviewer can check the claim rather than the drawing's confidence.

It is a fair question, and the answer is conditional. Coating and potting can let a design be assessed at a lower pollution degree, because the surface is no longer exposed to the contaminant the degree describes. That is a real relief in dusty or condensing installations.

But three conditions have to hold, and they fail quietly:

• The coating must be continuous over the whole creepage path, including the transition where a conductor leaves the potting and meets a terminal. That transition is where tracking starts.
• The material must be qualified for the electrical and thermal duty, not merely convenient for the process. A potting compound with a low tracking index reintroduces the problem it was bought to solve.
• It must be verified, not asserted. A design assessed at pollution degree 1 on the strength of a coating needs evidence that every production unit has the coating, at the thickness assumed, cured as assumed, with no voids at the critical transitions.

In practice the honest position for most laminated busbars is: declare the pollution degree the enclosure actually provides, and treat potting as a thermal and mechanical measure with an insulation benefit that is documented where it is real.

Take the drawing, take the four declared inputs, and walk these checks.

1. Write U_i, overvoltage category, pollution degree, material group and altitude in the title block. If any is missing, stop; the distances cannot be validated without them.
2. Look up the impulse level for the category and U_i, then the clearance for that level at your pollution degree. Apply the altitude factor if above 2,000 m. Compare with the smallest air gap on the drawing.
3. Look up creepage for the working voltage, pollution degree and material group. Compare with the smallest surface path — including at the laminate edge and at any terminal where a lug lands.
4. Mark the five shortest paths on the drawing, whichever direction they run. If you cannot mark them, the number on the drawing came from somewhere else.
5. Check the paths after hardware: bolts, washers, lugs, cable entry plates.
6. Decide, explicitly, which gaps are solid insulation and which are air. Write it down once, for the reviewer and for the next project.

Two hours of this on one drawing is usually enough to find the pattern, and the pattern is what gets reused for every variant.

Add these to the release checklist and most of the argument disappears before it reaches the customer.

• U_i, overvoltage category, pollution degree, material group stated on the drawing.
• Altitude stated, with correction applied when above 2,000 m.
• Minimum air clearance derived and written, not inherited from an old drawing.
• Minimum creepage derived and written, with the governing surface identified.
• Internal layer gaps classified as solid insulation, with the film system named.
• Laminate edge treatment specified — cap, potting, or minimum edge distance.
• Hardware included in the distance check: bolt heads, washers, lugs.
• Coating or potting claimed as insulation benefit backed by a process specification.
• Working voltage for creepage stated separately from U_i, with both values on the drawing.
• Any deviation from the table documented, with the reason and the approver.

If you want a second pair of eyes on a stack before it reaches a panel builder, our engineering team reviews drawings against these inputs as part of quoting. Start with a quote request, or compare layer constructions in the laminated busbar design guide.