“How tight can we bend this tube?” is usually answered by a tooling catalogue, a past job, or a trial bend. It should be answered by arithmetic first, because the limit is set by two numbers you already have: the strain the outside wall has to carry, and the elongation the material can actually deliver.
This guide gives the calculation, the material check that turns it into a decision, and the honest limits of both — including why a formula that says “impossible” is sometimes wrong, and what to do when the drawing asks for less than the tube allows.
Quick answer: The minimum bend radius is the tightest centreline radius the tube can take without splitting the outer wall or collapsing the section. Screen it with the outer-fibre strain, ε = 50 ÷ (CLR/D) in percent, and compare that with the elongation on the material certificate. If required strain is more than about 15% above the available elongation, expect cracking and change the design or the process before you cut metal.

What “minimum bend radius” means on a tube drawing
Three radii get confused in quotation, and the confusion costs money:
| Term | What it is | Where you see it |
|---|---|---|
| Inside radius (IR) | Radius of the concave face | Tooling drawings, some CAD |
| Centreline radius (CLR) | Radius to the middle of the wall — this is the number that governs bending | Tube drawings, bend tables, 1.5D / 2D callouts |
| Outside radius (OR) | CLR + half the OD | Occasionally on assembly drawings |
Tube bending is normally specified as CLR/D — the centreline radius divided by the tube outside diameter. A “1.5D bend” on a 38 mm tube means a 57 mm centreline radius. That single ratio appears in every calculation on this page, in mandrel selection charts, and in every defect prediction, which is why it is worth fixing the vocabulary early.
If your question is how to measure or calculate a radius rather than how small it can go, the bend radius measurement and calculation guide covers the geometry side, and bend allowance and developed length covers the flat-length arithmetic.
The two ratios that decide the whole job
Everything that happens in a bend follows from two dimensionless numbers. Neither depends on the machine.
| Ratio | Definition | What it controls |
|---|---|---|
| CLR/D | Centreline radius ÷ outside diameter | Outer-fibre strain, ovality risk, mandrel need, springback severity |
| Wall factor | Outside diameter ÷ wall thickness | Whether the wall can resist collapse between supports |
| Wall factor is sometimes written the other way up as t/D, which is the same information |
A part with a low CLR/D and a high wall factor is the hardest combination there is: tight bend, thin wall. A part with a high CLR/D and a low wall factor is the easiest. Before you discuss tooling, put the part in one of those two corners — it tells you whether you are arguing about a radius problem or a support problem.
The strain you cannot design out
When a tube bends, the material at the outside of the bend has to travel further than the material at the centreline. The extra distance is the strain, and no tooling choice removes it — tooling only decides how well the tube survives it.
For a bend where the neutral axis stays at the centreline, the elongation at the outer fibre is:
Outer-fibre elongation (%) = (OD / 2) ÷ CLR × 100
which is the same as:
This is geometry, not opinion. For a 38.1 mm tube on a 1.5D bend (CLR 57.15 mm):
One third of the material’s length at the outside of the bend has to stretch by a third of its own length. That number explains most of what goes wrong in tube bending, and it takes ten seconds to compute.
| CLR/D | Outer-fibre elongation required | Wall thinning at the same bend |
|---|---|---|
| 1.0 | 50.0% | 33.3% |
| 1.25 | 40.0% | 28.6% |
| 1.5 | 33.3% | 25.0% |
| 2.0 | 25.0% | 20.0% |
| 2.5 | 20.0% | 16.7% |
| 3.0 | 16.7% | 14.3% |
| 4.0 | 12.5% | 11.1% |
| 5.0 | 10.0% | 9.1% |
The thinning column is not a separate phenomenon — it comes from the same geometry and is covered in the next section.
The ductility check: the one test you can run before cutting metal
The required elongation is only useful if you compare it with what the material can deliver. That number is on the mill certificate, in the “elongation” or “elongation at fracture” field, usually quoted in percent over a 50 mm gauge length.
Decision rule:
| Situation | Reading |
|---|---|
| Required elongation well below the material’s | Radius is feasible; tooling decides the quality, not the fracture risk |
| Required elongation within about 15% of the material’s | Marginal — plan a trial bend and expect to control it tightly |
| Required elongation well above the material’s | The bend will crack or tear; change the radius, the material or the condition |
Typical published elongation values, to be replaced by the certificate for the actual lot:
| Material / condition | Typical elongation | Strain screen suggests a minimum around |
|---|---|---|
| 1018 / 1020 mild steel, annealed | 15–25% | 2.5D–3.5D |
| 304 / 316 stainless, annealed | 35–40% | 1.25D–1.5D |
| 6061-T6 aluminium | 10–12% | 4D–5D |
| 5052-H32 aluminium | 12–18% | 3D–4D |
| C122 copper, annealed | 40–50% | 1D–1.25D |
| Titanium Grade 2 | 20–24% | 2D–2.5D |
Note what the 6061-T6 row is telling you: on strain alone, T6 aluminium needs a radius around 5D. That matches the well-known warning in design guides that 6061-T6 is the outlier among common metals and needs several times the radius of mild steel. If a drawing calls for 1.5D in 6061-T6, the honest answer is that no tooling purchase solves it — the material has to be bent in the O condition and aged afterwards, or the design has to change. See CNC tube bending material selection for how material condition drives the whole process choice.
Worked example: three materials, four radii
Take a 38.1 mm OD tube, 1.5 mm wall, and ask what each material can take. The required strain depends only on CLR/D, so the same column applies to all three.
| CLR/D | Required strain | 304 stainless (40%) | Mild steel (20%) | 6061-T6 (12%) |
|---|---|---|---|---|
| 1.5 | 33.3% | Feasible | Marginal | Will crack |
| 2.0 | 25.0% | Feasible | Marginal | Will crack |
| 3.0 | 16.7% | Comfortable | Feasible | Marginal |
| 5.0 | 10.0% | Comfortable | Comfortable | Feasible |
Read the same row across and you have the quotation conversation: the stainless part is a tooling job, the mild-steel part is a process-control job, and the aluminium part at 1.5D is a drawing revision.

What crossing the limit actually looks like
Four different failures get called “the radius is too tight”, and only the first one is really about the bend radius. Reading the symptom correctly tells you which ratio you have run out of.
| What you see | What failed | Governed by | First move |
|---|---|---|---|
| Transverse crack or tear on the outer crown | Outer-fibre strain exceeded the material’s elongation | CLR/D | Larger radius, softer condition, or hot bending |
| Wrinkle, fold or bump on the inner radius | Inner wall buckled in compression | Wall factor and support | Mandrel ball count, wiper die contact |
| Section collapsed flat across the bend | Loss of section stiffness | Wall factor and internal support | Mandrel size and clearance |
| Split running along the tube axis | Weld seam positioned on the outer fibre | Seam orientation | Rotate the seam to the neutral axis |
| Angle or radius drifts off nominal | Springback, not a fracture limit | Material and CLR/D | Program compensation, then re-measure |
The practical value of the table is the second row. A wrinkle on the inner radius means the bend radius was achievable — the material survived the strain — and what failed was support. Buying a bigger radius to fix a wrinkle is the most common wasted design change in tube fabrication, because the same radius would have run clean with the right tooling.
Conversely, a transverse crack on the outer crown cannot be fixed with tooling, whatever the catalogue says. That is the strain limit, and the answer is geometry or material.
Why this screen is conservative — and when to overrule it
Applied strictly, the strain formula will tell you that a 1.5D bend in mild steel is impossible, and most shops have shelves of 1.5D mild-steel parts that prove otherwise. That is not a failure of the arithmetic; it is a limit of the assumption behind it.
The formula assumes the neutral axis sits at the centreline. It does not. As the bend gets tighter, the neutral axis shifts toward the inside of the bend, which shortens the path the outer fibre has to travel and lowers the real strain. This is the same effect sheet metal shops handle with a K-factor, which commonly falls between about 0.33 and 0.50 of the thickness, and it moves in the same direction in tube bending.
Three more reasons the real limit is more forgiving than the screen:
- Local strain capacity exceeds certification elongation. The elongation figure is measured to fracture over a gauge length and includes diffuse necking; a supported local bend can carry more before it tears.
- Internal support suppresses fracture initiation. A correctly sized mandrel changes the stress state at the outer wall, which is one of the reasons mandrel bending reaches radii that unsupported bending cannot. Whether a given part needs that support, and which mandrel arrangement fits it, is a tooling decision in its own right; the mandrel bending technology guide covers that decision.
- The wall thins as it stretches. Thinning relieves some of the constraint, and the thinning column in the table above is exactly what pays for the extra length.
So treat the result as a risk ranking, not a verdict. Where the screen says “marginal,” and your shop has bent that part for years, your shop is right. Use the formula to sort new jobs, to justify a trial bend, and to explain to a customer why their 1.2D drawing needs a conversation — not to overrule proven production.
Wall thinning comes from the same triangle
Because the outer wall has to cover more length with the material it has, thinning is not an independent defect — it is the strain budget being paid in thickness. The standard relation is:
where Ro is the outside radius, CLR + OD/2. That simplifies to:
Which gives the right-hand column of the table earlier in this page, and an interesting result: the widely quoted industry practice of 10–20% wall thinning corresponds exactly to bends between 5D and 2D. The rule of thumb was never arbitrary — it is the geometry, restated.
For the remaining wall rather than the percentage, and for how to measure it after the bend, see the wall thinning calculation and acceptance limits page.
What actually shifts the minimum radius
The strain is fixed by geometry. What changes is how much of that strain the tube survives, and four things move it:
| Lever | Effect on the achievable radius | Cost |
|---|---|---|
| Internal support (mandrel) | Largest single improvement, especially below 2D and on high wall factor | Tooling, set-up, maintenance |
| Wiper die | Extends the limit where the inner radius fails first | Tooling, set-up |
| Axial boost / pressure-die feed | Reduces thinning at the outer wall, allowing tighter bends | Machine capability, programming |
| Wall factor (OD ÷ wall) | Lower wall factor (thicker wall) tolerates tighter bends | Part weight, material cost |
| Material condition | Annealed or O-temper material bends far tighter than T6 or cold-worked | Post-processing, strength |
| Weld seam position | Seam on the neutral axis bends better than seam on the outer fibre | Fixture discipline, not cost |
The important ordering: tooling extends the limit; it does not remove it. A mandrel can take a 304 tube from “wrinkles at 2D” to “clean at 1.5D”. It cannot take 6061-T6 from 5D to 1.5D. When the gap is a factor of two in strain, the answer is material or design, not equipment.
For the tooling side in detail, see mandrel size, clearance and wear and the thin-wall tube bending process window, which covers the wall-factor limit separately from the strain limit.
The bend radius can be fine and the part still not makeable
There is a second minimum that has nothing to do with strain, and it catches more projects than fracture does: tooling clearance. A 1.5D bend in 304 may be perfectly achievable on a sample and impossible on the actual part.
| Constraint | What runs out | Practical guide |
|---|---|---|
| Straight between two bends | The clamp die and pressure die need room to grip and release | Commonly at least 1.5× to 2× the tube OD of straight between tangent points |
| Bend close to the tube end | The clamp die needs length to hold | The tail must exceed the clamp die length plus a margin |
| Total tube length | Machine bed, carriage travel, booster stroke | Check against the machine envelope, not just the bend table |
| Multi-plane bends | The bent leg can hit the machine head or the floor | This is an interference problem, and it is what simulation is for |
| Finished part handling | Very tight bends make the part hard to load and unload | Cycle time, not feasibility — but it still costs money |
The message is worth stating plainly to anyone quoting from a radius table: a radius is necessary but not sufficient. The part is makeable when the strain is survivable, the tooling fits, and the machine can reach. Check all three before the drawing is frozen.
For the interference question, see 3D tube bending and plane rotation and tube bending simulation.
A practical starting table
Combining the strain screen with the wall-factor limit and normal workshop practice gives a starting point. Treat every entry as a radius to open the conversation with, not a capability claim — the trial bend is what settles it.
| Material / condition | Practical minimum CLR/D | Binding constraint |
|---|---|---|
| Copper, annealed | 1.0D–1.5D | Tooling and surface finish |
| 304 / 316 stainless, annealed | 1.5D | Strain, then wrinkling |
| Mild steel, annealed | 1.5D–2D | Strain at thin wall |
| 5052 aluminium, H32 | 2D–3D | Strain |
| Titanium Grade 2 | 2.5D–3D | Springback and strain |
| 6061-T6 aluminium | 3D–5D (or bend in O, then age) | Material elongation |
| Any material, wall factor above about 40 | Add one D, or expect a mandrel and a wiper | Wall collapse |
How to prove the limit on your own tube
The screen ranks the risk; a trial bend settles it. Five steps, and the record matters as much as the result:
- Use the production lot. A sample from a different heat or a different mill can pass where production fails. Material condition is the single most common reason a “qualified” radius starts cracking.
- Bend with production tooling and production settings. A radius proved with a borrowed mandrel and different lubrication proves nothing about the job.
- Inspect the right things, at the right place. Wall thickness at four points around the apex, ovality at the apex and at both tangent points, the inner radius visually for wrinkles, and the outer crown under magnification or dye penetrant where the part is safety-related.
- If it fails, step the radius, don’t guess. Move in 0.25D increments and record the radius at which each characteristic passes. That gives you a number to quote, not an opinion.
- Write it down and re-qualify on change. Freeze the qualified radius with the material grade, condition, wall, tooling and settings. Re-qualify when the material condition, the tooling or the machine changes.
For pressure-containing or safety-critical parts, the acceptance method and any additional examination come from the governing specification, not from this page. Ultrasonic wall measurement is commonly specified using ASTM E797/E797M as the practice reference.
When the drawing asks for less than the tube allows
There are only four honest answers, and three of them are cheaper than they look:
- Increase the radius. The cheapest change and the one designers resist most. Going from 1.5D to 2D drops required strain from 33% to 25% — often the difference between cracking and not.
- Change the material condition. Bend in the annealed or O condition and restore strength afterwards — solution treat and age for 6061, or specify an annealed temper for stainless that will not be highly loaded.
- Change the process. Hot bending or induction-assisted bending raises the elongation available at the moment of forming, at the cost of scale, distortion and a slower cycle.
- Add support and accept the residual defects. A mandrel plus wiper plus boost gets you to the tooling limit, and the remaining ovality or thinning then either meets the drawing or it does not.
There is no fifth answer where the tube cracks and the part is still good. If a supplier promises a 1D bend in 6061-T6 with no concession on any of the four, ask to see the first-article measurements.
The four deformations, in one table
Everything above predicts the four things that go wrong. Each has its own page because each has its own fix — the point here is that they share one cause.
| Deformation | Where it comes from | Governed by | Full treatment |
|---|---|---|---|
| Wall thinning | Outer fibre stretching | CLR/D | Wall thinning calculation |
| Ovality / flattening | Section collapsing with no internal support | CLR/D and wall factor | Cross-section ovality |
| Wrinkling / collapse | Inner fibre compressing without support | Wall factor, tooling | Inner-radius wrinkling |
| Springback | Elastic recovery after load release | Material, CLR/D, wall | Springback causes and measurement and springback compensation |
For the diagnostic side — which defect points at which setting — see defect triage for bent tubes. For what to write on the drawing once the radius is agreed, see the tube bending tolerances guide.
Using this before you quote
The purpose of all of the above is to settle feasibility in a conversation rather than in a scrap bin. Send three things and the radius question usually resolves in one exchange: the tube OD and wall, the centreline radius or the CLR/D callout, and the material grade with its condition — plus the drawing if there are bends near each other or near the tube end.
With those, a supplier can tell you whether the strain is survivable, what tooling it needs, and whether the part is reachable on the machine — instead of promising a radius and finding out at first article. If you are assessing equipment at the same time, the CNC pipe bending machine range shows the machine side, and mandrel size, clearance and wear covers the tooling that decides whether the bend runs clean. To have a specific part reviewed, contact Zhuoran Machinery.
For the complete forming and inspection sequence, see the tube bending process guide.
Frequently asked questions
How do you calculate the minimum bend radius of a tube?
Divide the outside diameter by two, divide by the centreline radius, and multiply by 100. That gives the outer-fibre elongation in percent. Compare it with the elongation on the material certificate: if the requirement is well below the available elongation the radius is feasible, and if it is well above, the bend will crack. For tube bending, quote the result as CLR/D so it can be compared across tube sizes.
What is a 1.5D bend?
A bend whose centreline radius is 1.5 times the tube outside diameter. A 2-inch tube on a 1.5D bend has a 3-inch centreline radius. It is the common “tight bend” in tube fabrication, achievable in most annealed metals with a mandrel, and difficult or impossible in low-elongation material such as 6061-T6.
Is the minimum bend radius the inside radius or the centreline radius?
For tube bending the working number is the centreline radius, because that is what the strain calculation and the tooling both use. When a drawing says “bend radius” without qualification, confirm which one is meant before quoting — a mix-up between inside and centreline radius is a routine source of rejected parts.
Does a mandrel reduce the minimum bend radius?
Yes, substantially, and it is the largest single lever after geometry. A mandrel supports the tube internally so the section resists collapse and the outer wall is not asked to carry the whole deformation alone. It extends the achievable radius; it does not remove the strain limit, so it cannot rescue a material whose elongation is far below the requirement.
Why does 6061-T6 crack when 5052 does not?
Elongation. 6061-T6 in the T6 condition has roughly 10–12% elongation, while softer aluminium tempers run 12–18% or more, and the required strain at a tight radius is 25–33%. The arithmetic fails before the tooling is even discussed. Bend 6061 in the O condition and age afterwards, or use a larger radius.
What is the minimum bend radius for stainless steel tube?
For annealed 304 or 316, 1.5D is a normal working minimum with mandrel support, and it is close to the strain limit rather than the tooling limit. Cold-worked or high-strength stainless needs more. Always check the certificate rather than a table — the same grade in a different condition behaves differently.
How does wall thickness affect the minimum bend radius?
Two ways, and they are often confused. Thicker wall lowers the wall factor, so the tube resists collapse better and tolerates a tighter bend. But wall thickness does not change the outer-fibre strain at a given CLR/D — so a thick wall helps the collapse limit and leaves the fracture limit where it was.
Can a tube be bent tighter than 1D?
Rarely by rotary draw bending, and not at all in low-elongation material. Below about 1D the outer-fibre strain passes 50%, which no common structural metal will carry in a single draw. Tighter “bends” are usually produced by a different process — roll bending for a large-radius sweep on a closed shape, or forming operations rather than bending — or by accepting a much thicker wall and a great deal of iteration. If a drawing shows a sub-1D radius, treat it as a process question rather than a tooling purchase.
What is the minimum bend radius for 1 inch tube?
It depends on the material and wall, not on the diameter. A 25.4 mm tube on a 2D bend has a 50.8 mm centreline radius and needs 25% elongation; the same tube in annealed 304 will take it comfortably, while the same tube in 6061-T6 will crack. Use the strain calculation, then check the certificate — the inch dimension alone cannot answer the question, which is why tables listing “minimum radius by tube size” are only useful as a rough starting point.
Technical references
- Bend Tooling. “Tube Bending Calculations & Formulas.” Industry reference for the elongation (
Pe = (Ro/R) − 1) and wall-thinning (Pt = (Ro − R)/Ro) relations. bendtooling.com/tube-bending-formulas - Xometry. “Top Tube Bending Design Tips.” Publishes the required-elongation relation
[(Ø/2)/CLR] × 100. xometry.com - Atlas Manufacturing. “Sheet Metal Bend Radius Guide.” Notes that 6061-T6 requires several times the bend radius of mild steel. atlasmfg.com
- Arcus CNC. “Bend Radius in Sheet Metal Work.” Outer-fibre strain
ε = t/(2r + t)and the ductility basis of minimum bend radius. arcuscnc.com - Ozcan, B., Evis, Z., & Ozturk, F. “Effect of mandrel position on bending quality in the rotary draw bending process.” Proc. IMechE Part L, 240(2), 232–249. DOI: 10.1177/14644207251318362
- Chen, J., E, D., & Zhang, J. “Effects of process parameters on wrinkling of thin-walled circular tube under rotary draw bending.” Int. J. Adv. Manuf. Technol., 68, 1505–1516 (2013). DOI: 10.1007/s00170-013-4938-5












