UL 891 versus IEC 61439 for busbar fabrication

Panel builders exporting from an IEC market into North America usually discover the difference between these two standards in the wrong order. They engineer the assembly, verify it to IEC 61439, and then find that none of that work converts into anything a US inspector will accept. The reverse trip is no easier. The two documents cover comparable products and reach comparable safety outcomes by mechanisms that are structurally different, and the differences reach all the way down into the drilling and plating of the bar.

Where each one stops

UL 891 applies to dead-front switchboards nominally rated 1000 V or less, intended for installation under the Canadian Electrical Code Part I, the National Electrical Code (ANSI/NFPA 70) and the Mexican installations standard NOM-001-SEDE, with available short-circuit current up to 200000 A. It excludes theatre preset and dimmer control equipment, live-front construction, railway control and electrification equipment, and constructions intended only to receive motor control centre units.

IEC 61439 covers low-voltage switchgear and controlgear assemblies to 1000 V AC and 1500 V DC, applied as Part 1 (general rules) plus a product part. Part 2 covers power switchgear and controlgear assemblies, the closest counterpart to a UL 891 switchboard.

The categories are not congruent, and the mismatch matters commercially. UL 891 is one of several North American product standards partitioning a space that IEC covers with one family. Drawout low-voltage power circuit breaker construction falls under UL 1558, not UL 891. Industrial control panels fall under UL 508A. A single IEC 61439-2 assembly design can straddle two or three of those North American categories, and each category is a separate listing exercise. Deciding which one you are in is the first engineering decision on the project, not a paperwork detail at the end. The broader map is in which busbar standard applies.

Conformity assessment: the difference that governs everything else

The difference here is procedural rather than technical, and it drives everything downstream of it.

Under IEC 61439 the manufacturer declares the assembly's characteristics and holds a design verification file demonstrating that the declared values are met. Verification is by testing, by derivation from a tested design, or by calculation within the limits the standard sets. No third party is required. In the EU the declaration is made under the Low Voltage and EMC Directives with EN 61439 as the harmonised route. The competence, the evidence and the liability sit with you.

Under UL 891 the product is listed by a nationally recognised testing laboratory. The NRTL evaluates the construction, witnesses or conducts the tests, and then maintains the listing through follow-up inspection at your plant: unannounced visits checking that what you are building still matches what was listed. The mark on the door is not a statement about a test that happened once. It is a statement that a factory is under ongoing surveillance.

Three consequences follow that catch exporters out.

Your verification file does not transfer. IEC test data may inform an NRTL's evaluation, but it does not substitute for listing. Budget the North American route as a separate programme with its own schedule, from the beginning.

Changes are constrained differently. Under IEC you can revise a design and update your verification file on your own authority, within the derivation rules. Under a listing, a construction change outside the listed configurations means going back to the NRTL. A shop used to IEC flexibility will find this the biggest practical adjustment.

Field modification works differently again. A busbar tap added on site to an IEC assembly is the original manufacturer's problem to assess. The same tap on a listed switchboard can void the listing and trigger a field evaluation. This has direct fabrication consequences: design spare pad positions and tap provisions in at the factory, because retrofitting them later is disproportionately expensive.

Temperature rise, and the plating decision it drives

Both standards limit temperature rise. They pick very different numbers, from different reference conditions, for different reasons. This is the single difference that changes what you do to the copper.

IEC 61439-1 sets its limits in Table 6. Terminals for external insulated conductors are limited to 70 K, and the footnote is important: that 70 K is a value based on the conventional test, and where a built-in component's terminal is also the terminal for an external conductor, the lower of the component manufacturer's limit and 70 K applies.

For busbars themselves, Table 6 does not give a single number. The row reads that the limit is set by the mechanical strength of the conducting material, possible effect on adjacent equipment, the permissible temperature limit of insulating materials in contact with the conductor, the effect of conductor temperature on the apparatus connected to it, and, for plug-in contacts, the nature and surface treatment of the contact material. The 105 K figure everyone quotes lives in note g, as a ceiling on top of those criteria: assuming all the listed criteria are met, a maximum temperature rise of 105 K for bare copper busbars and conductors shall not be exceeded. Note 1 explains where 105 K comes from: it is the temperature above which annealing of copper becomes likely. It is a metallurgical limit, not an electrical one.

The reference ambient is a mean of 35 °C in service, with test ambient permitted anywhere between 10 °C and 40 °C.

North American practice starts somewhere else entirely. The limits derive from the ampacity of the devices and conductors connected, mandated through NEC Article 110.14(C). IEC 61439-1 tabulates them itself, in informative Annex M, and the two busbar values are the ones to memorise: 50 K for unplated busbars and 65 K for plated busbars. Terminals are limited to 50 K except in the cases the annex qualifies. The reference ambient in North American switchboard practice is 40 °C.

Set the two side by side. The IEC ceiling on bare copper is 105 K; the North American limit on the same unplated bar is 50 K. That is less than half, before any argument about reference ambient. Plating recovers 15 K of it.

Two fabrication consequences follow, and they point in opposite directions.

For a North American job, plating becomes a design requirement at any serious current density. The 15 K gap between 50 K and 65 K buys a substantial fraction of a bar size across a whole switchboard. Tin over copper is the common choice; silver is used where contact resistance stability matters more than cost.

For an IEC job, the usual advice runs the other way. The Copper Development Association's guidance is not to plate copper-to-copper joint faces unless the environment demands it, because soft plating can flow at elevated temperature and relax the contact pressure the joint depends on. That advice is sound for the IEC regime and wrong for the North American one, and the reason is not metallurgical disagreement. The two regimes are limiting different things. IEC is protecting the copper from annealing; North America is protecting the connected wire and devices from heat conducted out of the assembly.

If you build for both markets, plating is a per-project decision keyed to the destination, and it belongs on the routing card, not in a standing works instruction.

Spacings versus clearance and creepage

UL 891 specifies spacings: minimum distances through air and over surface, between parts of opposite polarity and from live parts to ground, looked up from a table indexed by voltage band. The values are in inches, the bands are discrete, and interpolation between them is not permitted.

IEC does not tabulate distances that way. Clearance and creepage in IEC 61439 are derived through IEC 60664-1: clearance from the rated impulse withstand voltage together with pollution degree and altitude; creepage from working voltage, pollution degree and the comparative tracking index of the insulating material.

The practical difference is what each system is sensitive to. Two IEC assemblies at the same nominal voltage can carry different minimum distances because one sits in a dirtier environment or at higher altitude. A UL spacing table does not ask those questions at that level. Some UL standards now permit UL 840, which is aligned to IEC 60664, as an alternative to the legacy table, but UL 891's own construction requirements are what your listing is assessed against.

For a fabricator the consequence is bar layout. A design laid out to IEC clearances at pollution degree 3 will usually satisfy UL spacings, and a design laid out to UL spacings at 600 V will usually satisfy IEC at pollution degree 2. Usually, in both directions, is not a design method. If a single physical design has to serve both, lay it out to the envelope of both, and accept the extra width.

This is also where a common vendor claim needs correcting: IEC 61439 does not mandate chamfering. There is no clause requiring an edge radius. Edge geometry is an engineering means of meeting the clearance, creepage and dielectric provisions, and it earns its keep because field stress concentrates at a sharp corner. It is a good idea for good reasons, and the reasons hold up when a customer's engineer asks.

Bracing and short-circuit withstand

Both standards require the assembly to survive fault current, and they characterise it differently.

IEC 61439 declares rated short-time withstand current (Icw) for a stated duration, rated peak withstand current (Ipk) and, where applicable, rated conditional short-circuit current (Icc). The test on main busbars is prescriptive about geometry: the length of main busbar under test is fixed at (2 ± 0,4) m and must include at least one joint where the busbars are intended to be extendable. The acceptance criteria are equally specific: clearances and creepage distances must still be met after the test, busbar supports must not have separated or cracked through, and there shall be no loosening of parts used for the connection of conductors.

UL 891 assigns a short-circuit current rating up to 200000 A, established with the specified overcurrent protective devices in place, and the rating of a switchboard section is limited to the lowest rating in the series path. Bracing is a marked construction attribute rather than a declared electrical parameter, and where cables routed through the switchboard needed bracing during the test, that requirement is marked on the equipment. Standard testing is short, a few cycles matched to the clearing time of a moulded-case device, with a longer duration required where a short-time rating is needed for selective coordination. UL 891 also provides a construction-rules route in its Annex G to a 100,000 A rating without a short-circuit test, which is the closest North American analogue to IEC's non-test verification options.

For a busbar shop the fabrication consequence is identical under both, which is the useful part: the joint is the weak point. IEC says so explicitly in its acceptance criterion, and UL's bracing philosophy says the same thing structurally. Bolt preload that has relaxed, a joint face that never made full contact, an elongated hole from insufficient edge distance: any of these turns a passing design into a failing sample.

What changes at the machine

• Hole patterns: European drawings work to metric patterns derived from DIN 43673-1. North American drawings work to the NEMA two-hole convention: ½-inch bolts on 1¾-inch (44.45 mm) centres, with 0.562-inch holes. Neither is convertible into the other by rounding. If you serve both markets, you carry both tool sets, and the tool station count on the machine becomes a commercial constraint. That is the argument for a 24-station processing centre such as the IMAC-CENTER 80, extendable to 36 stations, over a fixed-tool line that needs a changeover every time the destination market changes. The punching process page covers the tooling side.

• Joint preparation: under both regimes the joint decides the temperature-rise result, but the North American limits leave you far less headroom to absorb a poor one. Flat, freshly abraded faces at assembly, controlled torque with the nut factor matched to whether the threads are lubricated, and washers sized to spread load away from the hole. Under IEC you might survive a mediocre joint at 105 K; at 50 K on unplated bar in a North American switchboard you will not.

• Plating: destination-driven, not standing practice. Where it is specified, the bar has to leave your shop clean and burr-free, because plating over a burr traps the defect permanently.

• Edge and burr condition: a burr in a joint face holds the two bars apart and the contact area you calculated is not the contact area you get. This is why deburring belongs in the process route rather than in the fitter's hands, and why lines like the BND800-2 exist for bar from 0.5 to 50 mm thick and 100 to 800 mm wide.

• Labelling: the nameplate content differs, and so does who owns it. A UL listing mark is applied under the terms of the listing, with the field content specified by the standard. An IEC nameplate carries the manufacturer's declared characteristics, including the rated diversity factor, which has no North American counterpart at all.

Sizing the bar for two markets

Because the temperature-rise budgets differ so sharply, the same duty produces different bar in the two regimes, and the difference is not a small margin. Work it from the limits rather than from habit: establish the applicable rise, apply the correct reference ambient, then size from a ratings source with its correction factors intact. DIN 43671 with its ambient and arrangement corrections is the usual working reference, and our ampacity calculator applies those corrections rather than assuming reference conditions.

For the IEC side of the same calculation, the verification route you can use is bounded by rated current, and above 1600 A the calculation route closes entirely. That constraint is set out with the rest of the boundary conditions in our temperature-rise verification article.