Edgewise versus flatwise busbar bending

Flatwise and edgewise are not two settings of the same operation. They are different forming problems with different limiting failure modes, different tooling and, on most machines, different capacity ratings. The vocabulary varies by region: flatwise is also called the easy way, the flat bend or the level bend; edgewise is the hard way, the edge bend or the vertical bend. The mechanics do not vary.

A flatwise bend folds the bar across its thickness, so the small dimension governs. An edgewise bend folds it across its width, so the large dimension governs. Everything else follows from that one swap.

The section modulus is the whole story

For a rectangular bar of width w and thickness t, the elastic section modulus about the two principal axes is:

• Flatwise: Z = w·t² / 6
• Edgewise: Z = t·w² / 6

The plastic moduli carry the same ratio, w·t²/4 and t·w²/4, and it is the plastic value that matters once you are forming. Either way the ratio between them is simply w/t.

That single number is the multiplier on bending moment. Some concrete cases:

Read the width dependence carefully, because it is the part that surprises people. Flatwise section modulus is linear in width: double the width and you double the force. Edgewise section modulus goes with the square of width: double the width and you quadruple it. A 200 mm wide bar is not twice as hard to bend edgewise as a 100 mm bar. It is four times as hard.

Thickness behaves the opposite way round. Flatwise force goes with t², edgewise force only linearly with t. So a thin wide bar is the worst possible edgewise candidate on both counts at once, needing high force while offering the least resistance to the failure mode described next. If you want tonnage figures for a specific section, the busbar bending force calculator will work it from the geometry.

Stability sets the limit before force does

Machines are moment-limited. Workpieces are stability-limited, and in edgewise bending the workpiece limit arrives first.

On the outside of an edgewise bend the fibre sits half a bar width from the neutral axis rather than half a thickness. The outer-fibre engineering strain is:

ε = 1 / (2·(R/w) + 1)

At an inside radius equal to the bar width, that is 33%. At twice the width, 20%. At four times the width, 11%. Compare those against the elongation figures in the EN 13601 tempers, where R240 half hard is specified at 8% minimum and R290 hard at 4%, and it becomes clear why edgewise radii are quoted as multiples of width rather than multiples of thickness. Bend strain tolerance does exceed tensile elongation, for the reasons set out in the article on copper temper and minimum bend radius, but the margin is not large enough to make R/w below about 2 a safe general rule.

On the inside of the bend the problem is different and usually governs. The inner fibre is in compression along an arc, and the material carrying that compression is a strip only t thick and w deep. It is a slender plate loaded in its own plane, and slender plates in compression buckle out of plane. The critical stress falls with the square of the slenderness, so as w/t rises the inner edge becomes progressively easier to wrinkle. The visible result is a wavy or scalloped inner edge, sometimes with the whole section rolling over out of square.

That is why edgewise capacity ratings are so much tighter than flatwise ones. On the SMART-603CNC-S, flatwise bending is rated at 260 × 20 mm and edgewise at 125 × 15 mm. Both dimensions come down, not just the width, and the same pattern shows up across busbar bending equipment generally. The rating marks the point at which the tooling can no longer hold the section square through the bend.

Practical limits

Combining the strain arithmetic with normal shop practice gives a working envelope. Treat these as indicative starting points to be confirmed on a first article, not as normative limits.

Inside radius of at least 2× bar width for soft and half-hard copper, and 3× or more for hard temper. Width-to-thickness ratio below about 10 for reliable results without specialised support; above 12 to 15, expect inner-edge wrinkling regardless of radius. Bar width within roughly half the machine's flatwise rating. Hard temper in wide sections is the combination that fails most often, because both the outer-fibre strain limit and the buckling limit are tightening at the same time.

Springback also behaves differently. The plastic zone geometry, the constraint from the tooling and the effective R/t are all different from the flatwise case, so an edgewise offset cannot be derived from a flatwise one. Keep the two sets of entries separate in the springback compensation table.

Tooling and support

An edgewise bend has to be constrained everywhere the section can escape to, and the tooling that does it is not the tooling that bends flatwise.

The bar has to be clamped against both wide faces through the whole bend arc, not just at the entry. Unsupported wide faces let the section roll, and a rolled section produces a bend that is out of plane as well as out of angle.

The inner edge needs a follower or pressure block that travels with the bend and holds the compressed edge against buckling. This is the component that distinguishes a machine that can bend edgewise from one that merely has an edgewise mode.

The former needs a groove matched to the bar thickness. A universal vee that works across a thickness range flatwise will not hold an edgewise section, because the clearance that gives it flexibility is exactly the clearance the section needs to roll into.

Lubrication matters more than it does flatwise, because contact pressures on the former face are higher and the bar slides further across the tool during the bend.

The straight lead-in also has to be longer. Clamping length is a dimensional constraint on part layout, and it is a common cause of a design that is bendable in principle but not on the machine that will make it. Keep holes and lap joints clear of the bend zone; allowing at least 10 mm between a bend and a lap joint, and 20 mm between a bend and a support point, is normal panel-shop practice.

The design alternatives

When an edgewise bend does not fit the envelope, there are three ways round it, and they are not interchangeable.

Twist and flat bend

Twist the bar 90° about its own axis so the wide face rotates into the plane of the required direction change, then make that change as an ordinary flatwise bend. The forming is easy, the tooling is standard, the outer-fibre strain in the bend is governed by thickness rather than width, and the bar arrives at the destination already in the plane the terminal wants. The cost is length: a twist needs its own run of straight bar, typically at least twice the bar width and often more. The busbar twisting guide covers the geometry and the achievable angles, and twisting is a standard capability on busbar bending heads rather than a separate machine.

Jointed corner

Make the direction change with two straight bars and a bolted lap joint. There is no forming limit at all, so this works at any width and any temper. The cost is a joint in the current path, which means contact resistance, a torque specification, a maintenance obligation and an inspection point. Panel practice sizes the overlap so that contact area is at least five times the bar cross-section. A joint is also a thermal discontinuity, so it needs to sit where the temperature-rise verification can accommodate it.

Segmented or profiled plate

Cut the corner as a single shaped piece from plate, so the direction change happens in the plane of the material and no forming is required. This gives the lowest resistance of the three, no joint, and complete freedom over the corner geometry. The costs are material yield, since the offcut inside the corner is scrap, and a part that is no longer a standard section on the drawing.

Choosing between them

A decision rule that holds up in practice:

If the bar sits inside the machine's edgewise envelope and the drawing allows an inside radius of at least twice the bar width, bend it edgewise. It is the fewest operations and the fewest parts.

If the width or the width-to-thickness ratio is outside that envelope, but the route has enough straight length available before the direction change, twist and bend flatwise. This is the default for wide thin bars, and it is usually cheaper than people expect because both operations run on the same bending head.

If there is no length available for a twist, and the corner is a natural place for a joint anyway, use a bolted corner. A corner that already needs a tap-off, a transition between sections, or a service break is a corner that should have a joint in it regardless.

If the current density or the enclosure geometry rules out a joint, and there is no room to twist, profile the corner from plate.

The order matters. Each step down the list adds either parts, operations or scrap, so work down it rather than starting at the bottom.

Where this shows up in design review

Most edgewise problems are decided long before the bar reaches the shop. A routing drawn in 3D with a 90° corner in the wrong plane, on a 160 × 10 bar, at a radius of 100 mm, is not a fabrication problem to be solved with more tonnage. It is a geometry that no amount of force will produce without a wrinkled inner edge.

The cheap intervention is to check three numbers at design review: bar width against the machine's edgewise rating, inside radius against bar width, and width against thickness. Two of the three failing means the routing needs to change. Machines such as the EMAC-BB-H12 that read 3D and 2D geometry directly make this check easy to automate, because the bend list is already in the file and the capacity envelope is already in the control.