Copper temper and minimum bend radius

"Copper busbar, 100 × 10, 3 metres" is not a purchasable specification. It names a metal and a section and leaves out the one property that decides whether the bar will bend to the radius on your drawing or crack across the outer face on the first part. That property is temper, and it is a supply-chain field before it is a workshop problem: once the bar is on the rack, nobody can change it without a furnace.

What follows indexes bend-radius guidance to the temper codes that appear on purchase orders and mill certificates, rather than to thickness alone. Thickness matters, but it is the second variable.

What the EN 13601 codes mean

EN 13601 covers copper rod, bar and wire for electrical purposes, and for busbar work the grade of interest is Cu-ETP, designation CW004A. The standard identifies material condition two ways, and both appear on certificates.

The R designation is keyed to minimum tensile strength in N/mm². R220 means a minimum tensile of 220. The H designation is keyed to minimum Vickers hardness. H040 means 40 HV minimum. The two run in parallel, which is why you see them paired.

The tensile ranges are the standard's. The proof strength, elongation and hardness figures are the ones mill datasheets publish alongside these designations for Cu-ETP and should be treated as typical rather than normative; the Copper Development Association's busbar guidance quotes 50 to 55 N/mm² proof for fully annealed busbar-section copper and 170 to 200 N/mm² for half hard, which is lower at the soft end than the strip figures above.

Two things follow from the table and both matter more than the numbers themselves.

First, the chain is one-directional. Cold work raises tensile strength, raises proof strength, raises hardness and cuts elongation. Every property in the row moves together because they all come from the same dislocation density. You cannot buy hard copper that also has soft-copper formability.

Second, the elongation column collapses far faster than the strength columns rise. Going from R220 to R290 gains you about 40% on tensile strength and costs you around 88% of your elongation. That asymmetry is why temper selection is dominated by forming considerations rather than by strength.

The size effect nobody mentions

There is a caveat on hard temper that catches people out on large sections. The maximum strength obtainable depends on the cross-section, because the amount of cold reduction a mill can put into a bar is limited by how much area reduction the process allows. CDA guidance puts hard-condition busbar in the range 250 to 340 N/mm² depending on cross-sectional area, with the larger sections at the lower end.

So "hard" on a 30 × 5 bar and "hard" on a 200 × 12 bar are not the same material state, even when both certificates say R290. The large section will sit nearer the bottom of the band, will be slightly more forgiving to bend, and will spring back slightly less. If you are transferring a proven offset table from a small section to a large one, expect it to be conservative rather than wrong.

Minimum bend radius, flatwise

The published starting point is CDA's guidance for high-conductivity copper in the half-hard or hard temper, which gives minimum former radii as a multiple of thickness:

Above 50 mm, bar is not normally bent at all; CDA notes it can be done with localised annealing before forming.

Those figures already assume the harder tempers, so they are a safe floor for R240 and R290. For R220 you can go tighter, but there is rarely a reason to. What the thickness bands do not show is the temper axis, so in practice a working table looks more like this:

Treat the R220 column and the R290 column as indicative. The CDA table is the verified anchor; the spread either side of it reflects normal shop practice and should be confirmed on a first article for each combination of material, thickness and tooling.

Elongation ranks tempers; it does not budget bend strain

The engineering strain at the outer fibre of a bend follows directly from the geometry:

ε = t / (2R + t) = 1 / (2·(R/t) + 1)

At R/t = 1 that is 33%. At R/t = 2 it is 20%. At R/t = 4 it is 11%. Compare those against the elongation column: R290 is specified at 4% minimum, which by naive reading would demand R/t above 11 before the outer fibre survives.

It does not, and knowing why stops people over-specifying radii. A tensile test measures elongation to fracture in a uniform gauge length where a neck is free to form and run. In a bend, the peak strain exists only at the outer surface over a short arc, and the surrounding unstrained material constrains any neck from developing. Achievable bend strain therefore exceeds tensile elongation by a wide margin, which is exactly how CDA can specify 1 t for hard copper up to 10 mm thickness on a material with 4% elongation.

Use the elongation figure to rank tempers against each other and to predict which will crack first. Do not use it as a bend-strain budget.

What failure looks like

Cracking at the outer fibre progresses through recognisable stages, and catching it at the first stage saves the batch.

Orange peel comes first. The outer surface roughens into a visible texture as individual grains deform independently. It is cosmetic on its own, but it tells you the outer fibre is close to its limit, and it worsens with coarse grain size, which is itself a symptom of over-annealing.

Transverse micro-cracks follow, running parallel to the bend axis across the outer face. On a plated bar they may only show as breaks in the plating. The part is scrap at this point, because the crack is a stress raiser sitting exactly where fatigue loading is highest in service. Through-cracking is the terminal stage and needs no description.

On the inside face the failure mode is different: compressive instability, showing as buckling or wrinkling. On a flatwise bend it is usually a sign that the bar was inadequately supported. On an edgewise bend it is the governing constraint rather than an incidental one.

Edgewise bending needs radically larger radii

Everything above is for flatwise bending, where the bar folds across its thickness and the small dimension governs strain. Bend the same bar the hard way, across its width, and the geometry changes completely: the outer fibre now sits at half the bar width from the neutral axis instead of half the thickness, and the inner fibre has a long unsupported span in compression.

The practical evidence is in machine capacity tables. The SMART-603CNC-S is rated at 260 × 20 mm for flatwise bending and 125 × 15 mm for edgewise, and that roughly halving of both dimensions is a pattern you will find across busbar bending equipment generally. Minimum radii for edgewise work run in multiples of bar width, not thickness, and are commonly several times the flatwise figure. The comparison of edgewise and flatwise busbar bending covers the mechanics and the design alternatives.

Do not carry a flatwise radius rule across to an edgewise bend. It is the fastest route to a cracked outer edge.

Grain direction

Rolled and drawn copper carries a crystallographic texture aligned with the working direction, and formability is directional as a result. The general rule from sheet forming applies: a bend whose axis runs perpendicular to the rolling direction is the favourable orientation, and a bend whose axis runs parallel to it is the one that cracks.

For ordinary busbar this works in your favour without any effort. Bar is rolled or drawn along its length, and a normal transverse bend puts the bend axis across the bar, perpendicular to the working direction. You are already in the good orientation.

Where it stops being automatic is bar cut from wide plate or slit from coil, where the length of the finished bar may not align with the original rolling direction. If a supplier substitutes slit material for drawn bar, the bend behaviour can change even though the certificate reads identically on strength. That is a good reason to record material form, not just grade and temper, in the bending process record.

Reading the mill certificate

A useful certificate for busbar carries the grade and standard, Cu-ETP CW004A to EN 13601. It carries the temper designation in both forms where the mill quotes both, R240 / H065. It carries measured tensile strength, elongation and hardness against the cast or lot number. It carries conductivity as a percentage of IACS, which for annealed material should be 100% or better and for hard-drawn material at least 97%, because cold work costs 2 to 3% conductivity.

What it usually does not carry, unless you ask, is measured 0.2% proof strength. That is the number the springback calculation needs, and for R220 it carries the widest permitted spread. The standard sets a ceiling on proof strength for soft material, not a floor, so two conforming R220 bars can differ by a factor of two in how far they spring back. If you run soft copper in production, put measured Rp0.2 on the purchase order.

The other thing to check is that the certificate refers to the delivered lot rather than to a type test. Type-test certificates are common and are not the same evidence.

When annealing is the right answer

There are legitimate cases. Bar over 50 mm thick, where localised annealing at the bend before forming is the only practical route. Rework of an already-formed bar, where the bend zone has been work-hardened by the first operation and a second bend at the same location would crack. Very tight radii on a retrofit where the bar is already cut and the alternative is scrap.

There is also a common case that is not a legitimate reason: buying hard bar because that is what the merchant had, then annealing it to make it bendable. That sequence pays for cold work, then pays again to remove it, and it discards the mechanical properties the hard temper was providing. Hard bar resists sag over long unsupported spans and resists permanent distortion under short-circuit electromagnetic forces. Anneal the whole bar and you have bought neither the formability of R220 nor the stiffness of R290.

Two technical constraints apply before you commit to annealing. Cu-ETP softens progressively above about 150 °C, and a full recrystallisation anneal is typically run around 400 °C, where recrystallisation is essentially complete within minutes. Longer soaks coarsen the grain, and coarse grain gives orange peel on the next bend, so a long anneal can make the surface finish worse while making the material softer.

The second constraint is specific to this grade. Cu-ETP contains residual oxygen. Anneal it in a hydrogen-bearing atmosphere and the hydrogen diffuses in, reduces the oxide inclusions and forms steam at the grain boundaries, which embrittles the material irreversibly. EN 13601 carries a freedom-from-hydrogen-embrittlement requirement for exactly this reason. If annealing is part of your process, control the furnace atmosphere or specify an oxygen-free grade instead.

Writing the purchase order

The minimum set for a busbar order is grade and standard, temper designation, section, form, and the certificate requirement. Adding measured proof strength costs the mill nothing and removes the largest single source of drift in bending. If a bar is going onto a bending machine with a stored springback offset table, the temper code is what that table is keyed on, and an unspecified temper means the table has no valid entry.

Temper also has no bearing on current-carrying capacity, which is set by section, arrangement and thermal environment; if that is the constraint you are working to, the busbar ampacity calculator handles it independently of how the bar was worked. The two decisions are separable, and keeping them separate stops formability arguments from leaking into sizing arguments.