Mandrel exhaust tube bending is the forming of thin-wall tube into the routed shapes that carry gas from a manifold to a tailpipe, holding the bore round enough that flow and joint sealing are unaffected. It is a rotary draw operation almost everywhere it is done in volume, and the difficulty comes not from the diameter but from the ratio between that diameter and a wall that is often thinner than a coin.
This article is about the process and about what a part imposes on equipment. Exhaust tube bending, and mandrel exhaust tube bending in particular, is written about here for people making exhaust components as a product, such as tier suppliers, aftermarket manufacturers, silencer fabricators, and builders of agricultural and generator equipment. It is not a guide to modifying a car. Machine pages should be read after you know what the part demands and not before, and the ranges are on the CNC automatic pipe bending machine and NC semi-automatic tube bender pages.
Why Mandrel Exhaust Tube Bending Is Harder Than the Size Suggests
A 60 mm structural tube with a 4 mm wall is an easy bend. A 60 mm exhaust tube with a 1.5 mm wall, bent to the same angle on the same radius, can collapse. The diameter did not change. The ratio did.
The diameter to wall ratio is the real difficulty index. Divide the outside diameter by the wall thickness. A Φ60.3 × 1.5 mm exhaust tube gives 40, and a Φ60.3 × 4 mm hydraulic line gives 15. As a working rule, tube above a ratio of roughly 20 needs internal support to hold its section through the bend, and essentially all production exhaust tube sits well above that line. This single number is why exhaust work is a mandrel discipline, and why quoting it against the headline maximum diameter of a machine tells you almost nothing.
The radii are tight because the vehicle is full. Exhaust routing has to thread past a floor pan, a driveshaft, a fuel tank and a rear axle, and the space left over sets the radius. Where a general fabrication drawing might specify 3D out of habit, exhaust parts routinely call for 1.5D and often 1.0D. Tight radius and thin wall are individually manageable, and together they are the whole problem.
There are also many bends in one part. An exhaust section commonly carries several bends, in more than one plane, with short straights in between. Each bend has to be positioned relative to the last within a tolerance that the downstream welding fixture will accept, and errors accumulate along the part and do not cancel out.
Finally, the bore is a functional surface. Unlike a chassis tube or a handrail, the inside of an exhaust pipe does work. Crushing the section at a bend restricts flow and raises backpressure, and a wrinkled bore disturbs it. That separates exhaust from most other automotive tube bending, where the tube is a structural member and only the outside geometry matters.
Five Readings That Decide the Exhaust Pipe Bending Machine
Everything above collapses into five readings. Take them in this order, because each one narrows the field before the next.
| Read this off the part | It decides | Why |
|---|---|---|
| Outside diameter and wall thickness | Machine capacity class | Capacity is rated as a diameter and a wall together, because bending torque depends on both |
| Tightest centerline radius | Tooling set, and whether a mandrel is mandatory | Below about 2D on thin wall, a mandrel stops being optional |
| Number of bends and plane rotations | NC or CNC control | Manual repositioning between bends is workable at two or three bends and not at many more |
| Shortest straight between two bends | Clamp die length and tooling layout | The machine needs somewhere to grip, and the part has to give it that length |
| Material and temper | Springback compensation, mandrel and lubrication | Two tubes of identical geometry in different alloys do not spring back the same |
Note what is not on that list, which is the maximum bending diameter of the machine. It is the number most enquiries lead with and the least useful of the set. A machine rated to Φ76.2 × 3 mm in mild steel is not automatically the right machine for thin-wall stainless at 1.2D, because the limiting factor there is not force but the tooling and mandrel set. Capacity ratings describe what the frame can push. The five readings describe what the job needs.
The Wall Thinning Budget in Exhaust Tube Bending
When tube is bent, the outside of the bend stretches and the inside compresses. The stretched material has to come from somewhere, and it comes from the wall. Thinning is not a defect to be removed. It is a consequence of geometry that has to be budgeted at the design stage.
A first approximation treats the neutral axis as the centerline. Thinning is roughly OD ÷ (2 × CLR + OD). For a Φ60.3 mm tube on a 150 mm centerline radius, a 2.5D bend, that gives about 17%. Tighten to 1.5D and it rises to 25%. Go to 1.0D and the geometry alone accounts for a third of the wall.
Two things move the real number away from that estimate. The neutral axis does not stay on the centerline. It migrates toward the inside of the bend, which reduces thinning below the geometric value. And a following pressure die, which advances with the tube and does not drag against it, feeds material into the bend and can cut thinning substantially. A tube bent with a correctly set pressure die and a well-fitted mandrel at 1.5D will measure noticeably better than the 25% the formula predicts.
On UK and Australian drawings the radius is called the centreline radius. What matters commercially is that the drawing states an acceptance limit. If it carries no number, the supplier is guessing, and the first shipment is where you find out whose guess was different. The way radius, wall and neutral axis interact is covered in centerline radius explained for tube bending, and the failure modes in tube bending defects and what causes each one.
Ovality Is the Feature That Actually Gets Rejected
Thinning is the number engineers discuss. Ovality is the number that fails parts at incoming inspection, because it is visible, easily measured and it stops joints sealing.
Ovality is the flattening of the round section at the crown of a bend, expressed as (Dmax − Dmin) ÷ Dnominal × 100. Two caliper readings at 90° across the crown are enough to check it. It is the cheapest quality check in the whole process and the one most often skipped.
The acceptable value depends on the part. Where a bend sits close to a slip joint, a sealing ring or a flange, the limit is tighter, because an out-of-round tube end will not seat evenly against a gasket however good the clamp is. Where a bend is in the middle of a straight run and simply carries gas, more is often tolerable. State the limit on the drawing and say where it applies, and do not assume a shared understanding of the word round.
Ovality is controlled by tooling and setup and not by machine size. The mandrel supports the bore through the tangent point. The pressure die holds the tube against the bend die along the straight. The wiper die fills the gap at the inside tangent. If the mandrel setback is wrong by even a couple of millimeters in either direction, ovality climbs on every part until someone corrects it. It is worth measuring on the first article and again after any tooling change.
Tooling Decides More of Mandrel Exhaust Tube Bending Than the Machine Does
A capable machine with the wrong tooling produces scrap, and mandrel exhaust tube bending shows it faster than any other job. This is the most common misallocation in exhaust enquiries, which is a generous budget on the machine and an afterthought on the die set.
Rotary draw bending uses up to five tools working together, and the tighter the radius the more of them you need. What matters for exhaust specifically is which of the five you can do without.
| Tool | Job | Needed on exhaust work? |
|---|---|---|
| Bend die | Forms the radius, and the tube is drawn around it | Always, one per radius unless stacked |
| Clamp die | Grips the tube to the bend die so it is drawn and not pushed | Always |
| Pressure die | Holds the tube into the bend die along the straight | Always, and following on tight radii |
| Mandrel | Supports the bore from inside through the tangent point | Effectively always at exhaust wall thicknesses |
| Wiper die | Fills the gap at the inside tangent to stop wrinkles forming | From roughly 1.5D and tighter, or on very thin wall |
Ball count follows radius. A plug mandrel with no balls will hold a gentle bend on heavier wall. Exhaust work generally needs a ball mandrel, and the number of linked balls rises as the radius tightens, with one or two at moderate radii and more at 1.0D to 1.5D on thin wall. Each ball extends bore support further around the arc, at the cost of more drag and more demand on lubrication.
Mandrel position is a setup parameter and not a fixed value. The nose sits slightly ahead of the tangent point. Too far back and the tube collapses because the bore is unsupported where it matters. Too far forward and you get humping on the outside of the bend and heavy drag on extraction. It is set to the millimeter and re-checked after tooling changes. The full comparison of tools is in tube bending tooling and how to match a die set to the job, and the process itself is on the mandrel tube bending page.
Telhoo’s published pages describe this pairing directly. The 75NC page says that thin-wall stainless exhaust wants a mandrel and a wiper die to stop wrinkling on the inside radius, and the 75CNC page describes matched wiper dies as what keeps thin-wall stainless from wrinkling in commercial vehicle and machinery exhaust.
Tooling is per radius, and that has planning consequences. A bend die is cut for one centerline radius. If a part carries a 1.5D bend and a 2.5D bend, that is two die sets, and on a single-stack machine it is a changeover between them. Multi-stack heads carry two or more radii on the same spindle, and on multi-radius exhaust parts they turn a changeover into an axis move.
The practical instruction for anyone drawing an exhaust part is to standardize the radius wherever the packaging allows. A part that uses one radius throughout needs one die set and one setup. A part that uses three, because three different people drew three sections, needs either a stacked head or three changeovers per batch for as long as it runs. The tooling decision is made in CAD, months before anyone quotes a machine.
Reading the Control Requirement in Mandrel Exhaust Tube Bending
The choice between semi-automatic and full CNC control is usually presented as a budget question. It is better read as a question about how much of the part the machine has to remember.
Every bend in a routed part is described by three values, often written as the length, rotation and angle. They are the straight length fed before the bend, the rotation about the tube axis that sets the plane, and the angle of the bend itself. On a semi-automatic machine the operator handles feed and rotation by hand or against stops, and the machine controls the angle. On a CNC machine all three are axes under program control, and the entire part is one program. The arithmetic that links them to the cut length is in pipe bending calculations.
| Part and batch pattern | Sensible control level | Reasoning |
|---|---|---|
| One to three bends, one plane, small batches, frequent changes | NC semi-automatic | Manual positioning is quick enough that automation earns little back |
| Many bends in several planes, repeating batches | CNC | Cumulative positioning error and cycle time both grow with bend count |
| Mixed radii within one part | CNC with a multi-stack head | A radius change becomes an indexed move and not a changeover |
| Left and right hand variants of the same part | CNC with bending in both directions | One program set and one tooling set, with no mirrored fixture |
| Prototype and low-volume aftermarket work | NC, or CNC if programs are reused | The value of CNC is repeatability across parts, which needs volume to pay |
Repeatability matters more here than on most tube work because of how exhaust parts are assembled. A finished system is several bent sections welded together in a fixture. If each section drifts a little, the fixture is where the drift shows, as gaps at the weld joints or as a system that no longer clears the floor pan. Telhoo’s CNC benders publish ±0.1° on angle and ±0.1 mm on feed, and the 75CNC lists servo axes with rotation of 25 to 180°/s and a 190° maximum bend.
A Worked Example of Reading an Exhaust Part
The five readings are easier to see on a part. This example is illustrative and is not a customer drawing. Take a stainless exhaust section of Φ60.3 × 1.5 mm with four bends, three at 1.5D and one at 2.5D, and a shortest straight of about one tube diameter between two of the bends.
The first reading gives a wall factor of 40, so the tube is far above the line where internal support is optional. Mandrel exhaust tube bending is the only sensible route. The second reading gives a tightest radius of 1.5D, which is 90.5 mm on this tube, so a ball mandrel with several balls and a wiper die are both on the list. The first estimate of outside thinning at that radius is 25%, so the drawing needs a stated limit and the process needs a following pressure die.
The third reading, four bends in more than one plane, points to CNC control, and the fourth, two radii in one part, points to a two-stack head so that the 1.5D and 2.5D dies are both on the machine. The shortest straight of one diameter is the one to check against the clamp die length before anything is cut, because a grip shorter than the straight leaves no room for the clamp. The fifth reading, 304 stainless, says that springback and lubrication need real attention and that the first article sets the compensation.
Run in that order, the five readings turn a vague request for an exhaust machine into a specific one. The tooling is a bend die for each radius, a clamp die long enough for the straight, a following pressure die, a ball mandrel with several balls and a wiper die. The machine is a CNC bender in the class that carries the diameter, with a stacked head. None of that needed the maximum diameter of the machine to be known first.
First Article Checks for Mandrel Exhaust Tube Bending
In mandrel exhaust tube bending, a first article is where the drawing meets the tooling, and a short routine catches most of the problems before a batch is run. Measure ovality at several points around each bend and not only at the crown. Section one bend and measure the remaining wall on the outside against the limit on the drawing. Look along the inside of each bend under a low angle light for the shadow pattern of early wrinkling. Check the angle after release against the programmed angle, then store the compensation against the material lot. Finally, drop the part into the checking fixture, because per-bend accuracy only matters if the assembled system closes.
Repeat the routine after any tooling change and at the start of each new material lot. It costs minutes, and it is the only reliable way to separate a tooling fault from a material change when something drifts.
Material Changes the Numbers
Mandrel exhaust tube bending runs on a small group of materials, and each one moves the process settings. The geometry stays identical and the compensation does not.
| Material | Typical use | What it changes at the bender |
|---|---|---|
| Aluminized carbon steel | Volume mid-sections | Bends easily with the least springback, but the coating scores if tooling or lubrication is wrong |
| 409 and 439 ferritic stainless | Standard long-life exhaust grades | Moderate springback, forgiving and predictable |
| 304 austenitic stainless | Performance and high-end work | Work-hardens and springs back more, so it needs more compensation and better lubrication |
| Titanium and nickel alloys | Motorsport and specialist parts | Narrow forming window, which is process development and not a catalog setting |
Springback is the practical expression of this. Release the clamp and the tube opens up. The machine deals with it by over-bending, and the first article establishes the actual figure while the program carries the compensation from then on. What matters when specifying is that springback is a property of your material batch and not of the machine. A change of supplier or coil temper can move it enough to need re-qualification, and a machine that cannot store a per-part compensation value turns that into an operator skill problem instead of a program parameter. The comparison across alloys is in bending stainless steel tubing.
The other material consequence is lubrication. A ball mandrel dragging inside thin stainless at a tight radius generates real friction, and dry running scores the bore, picks up material on the mandrel and eventually tears the tube. Lubrication of the mandrel is part of the process on a production exhaust line and not an accessory.
The Operations Either Side of the Bend in Exhaust Tube Bending
The bender is one station in a sequence, and exhaust parts fail as often at the boundaries as in the bend itself.
Before the bend comes the blank. Feed length accuracy is meaningless if the blank length varies. A saw cut that wanders puts that error into the first straight of every part, and burrs left in the bore will score a mandrel. An automatic pipe cutting machine holding a consistent length and a clean square end removes a whole class of argument about whose tolerance failed.
After the bend come the ends. Most exhaust joints are made by expanding one tube end to slip over another, or by flaring for a flange. That is a forming operation on a tube end forming machine, and Telhoo’s SG80-2 page names exhaust manufacturing as one of its main uses. The wall thinning already spent in a nearby bend helps decide whether the end forms cleanly or splits, so bends placed too close to an end form narrow the process window for reasons that have nothing to do with the bender.
Where sections are welded, the edges matter. A square, deburred and where required beveled end gives repeatable weld penetration, and a ragged one does not. That work belongs to a pipe chamfering machine and takes seconds per part.
When a part is rejected, the cause is frequently in this chain and not at the bend head.
What to Send for a Mandrel Exhaust Tube Bending Enquiry
Most exhaust enquiries arrive as a request for a machine for two inch exhaust pipe, which is not enough information to answer honestly. The list is short, and every item changes the recommendation.
- Tube size as OD × wall for every size you run and not just the largest. The smallest is often the harder job.
- Material and grade, and whether that is fixed or may change.
- Every centerline radius on the part, and whether the design is free to standardize on one.
- Bend count per part and the plane rotations, ideally as the length, rotation and angle table or the 3D model.
- The shortest straight between two bends, and the distance from the last bend to the tube end.
- Acceptance limits for wall thinning and ovality, with the tighter values called out where they apply.
- Annual volume and batch structure, meaning how many parts, how many variants and how often they change.
- What happens next to the part, such as end forming, welding into an assembly or coating.
Given those, a supplier can tell you the capacity class, whether a wiper die is required, how many mandrel balls the tightest bend needs, whether a stacked head pays for itself and whether the job wants NC or CNC control. Given only a diameter, they can tell you a price, which is a different thing. Send us the exhaust part drawing, and for a wider view of where this fits among other industries see tube bending applications by industry and automotive tube bending by part family and volume.
