How to obtain a stable tube bevel for subsequent welding ART-199

Start with the welded joint

First define the joint type, material, thickness, welding position and process, filler material, access, expected distortion and finished-assembly requirements. The required edge geometry follows from the engineering documentation and approved welding procedure, not from a universal table in this article.

The same external angle may behave differently with another root face or root gap. A tube-to-tube joint depends on mutual fit; a tube-to-plate joint depends on contact and access; a branch connection depends on the spatial contour around the perimeter. A bevel convenient for manual welding is not necessarily suitable for a robotic torch.

State in advance which parameters are drawing characteristics, which are controlled by the welding process, and which may remain cutting-process variables. This prevents the laser department from improving visual appearance while changing the functional joint geometry.

Bevel geometry has several parameters

Measure bevel angle from an agreed reference surface or local normal. The root face defines the material remaining at the root. The root gap appears between two parts and depends on both edges, the fixture and actual part dimensions. Root-face position describes where that land lies relative to the outside and inside diameters.

Variation around the tube perimeter also matters. An average angle of 35 degrees is not acceptable if one sector deviates substantially and creates a local root failure. In a spatial joint, parameters may intentionally vary; in that case, use a nominal-versus-actual map along the contour.

ART-198 addresses the general control of a square cut edge. ART-199 considers an intentional welding bevel. ART-182 answers the investment question of when a 3D head is justified. The present focus is the stability of a process that has already been selected.

Why the programmed angle is not the actual angle

The program defines the beam-axis path and head position, but the material has thickness and curvature. Beam path, kerf width, focal position and melt removal shape the outside and inside edge limits differently. On a sloping edge, the effective path through the material becomes longer, while gas access and melt ejection change.

If the tube is oval or its centre is displaced from the nominal axis, the local normal differs from the calculated one. A sensing system that maintains stand-off still does not guarantee correct bevel geometry: coordinate transformation and a validated cutting process are also required. Regripping or support movement introduces additional variation.

The finished angle therefore cannot be declared from the CAM value alone. Measure the part in the agreed datum system. The difference between commanded and measured angle becomes part of the process-capability study.

Kinematic access zone of the 3D head

Head inclination is limited by mechanical travel, collision with the tube or chucks, nozzle length, stand-off, access to the inner part of the contour and cable routing. An angle available at the top of the tube may not be available close to a chuck or on the opposite side of a complex section.

BLM GROUP and TRUMPF materials show the class of systems capable of complex tube processing. The exact model, head option, software licence and process package must still be verified separately. A 3D label is not a universal guarantee across every diameter, wall thickness and contour.

In the test matrix, include angular positions, proximity to the chuck, maximum inclinations and transitions between positive and negative bevel. CAM simulation should check collisions, but a physical dry run and first-part inspection remain necessary under the manufacturer’s procedure.

The incoming tube is part of the result

Measure diameter, ovality, wall thickness, seam, longitudinal bow and local dents. For sections, add corner radii and twist. Seam orientation can affect local cutting behaviour and subsequent welding. ART-193 discusses seam orientation in detail, but the factor must still be recorded in this test.

Compare several bars from typical deliveries. One straight coupon confirms only the result on that coupon. If production variation is wider, bevel repeatability may disappear. Material limits must be included in the accepted process envelope.

Do not straighten a sample before cutting by a method that is absent from normal production. Record support positions, chuck force, orientation and free length. Shape under clamping may differ from shape after unloading.

A process window without unsafe universal recipes

Quality depends on power, speed, focus, gas type and pressure, nozzle, piercing strategy and thermal sequence. The optimum for a square edge is not necessarily suitable for a bevel. Publishing universal numbers without a specific source, head and material would be unsafe.

Authorised specialists take the starting parameters from the manufacturer’s application tables and refine them through controlled cuts. Interlocks are not bypassed, unsuitable nozzles are not used and service limits are not exceeded. Every change is assessed against the full set of characteristics, not only speed or absence of dross.

Build a controlled process window. For every validated combination, retain material, wall thickness, diameter, angle range, contour type, gas, nozzle, software and measured output. This is a governed library rather than an arbitrary collection of “best settings.”

Test-coupon plan for a bevel

Create a coupon with a straight bevel, an angle transition, a contour in several angular positions, an end joint and a real spatial joint. Add repeats at different positions along the tube and after a changeover. If production covers several diameters and walls, include minimum, typical and near-limit combinations.

For every feature, record commanded and measured angle, root face, top and bottom edge positions, the agreed roughness indication, dross, heat tint, continuity and any finishing required. Measure mating fit and the root gap separately in the fixture.

Do not confuse a research coupon with a welding qualification coupon. The first investigates cutting capability; the second confirms acceptability of the complete welding procedure under company rules and applicable requirements.

Measuring angle and root face

A manual bevel gauge can support a quick shop-floor check, but its contact geometry on a round surface needs appropriate tooling. An optical profilometer, section microscope, coordinate-measuring machine or 3D system can provide a fuller profile. Select the method according to tolerance and available measurement uncertainty.

Define the portion of the edge used to fit the reference line. Micro-irregularities, dross and transition zones must not change the result accidentally. For destructive sectioning, record the section position and do not extrapolate one section to the whole circumference.

Perform a gauge repeatability and reproducibility study, or at least a repeatability check on representative coupons. If two operators systematically obtain different angles, process capability cannot be assessed reliably. Retain the raw profiles, not only a pass result.

Root face and inner edge

The root face often determines assembly stability, yet it is harder to inspect, especially inside a small tube. The outside bevel can look identical while the inner edge differs significantly. For a branch or saddle cut, access to the inside changes along the contour.

Use sectioned coupons, a replica method or a validated optical approach. Compare the outer and inner limits. If the design allows a different root face around the circumference, that variation must be specified rather than occur randomly.

After deburring, preserve traceability to the initial geometry. Measure before and after, or define the exact condition in which the part is accepted. Manual grinding can hide cutting instability while adding its own variation.

Thermal sequence and distortion

Long or repeated contours heat a thin wall. If a bevel covers a large sector, the local shape may change before cutting is complete. Start and end points, feature order, pauses and support influence the result.

Observe not only the edge but also ovality, opening and spring-back after unloading. ART-196 addresses thin-wall tube distortion in detail; ART-199 considers it only as a factor in bevel stability.

Compare the first, middle and last part of a batch. If the trend follows heat accumulation, one cold-start coupon does not represent the production run.

A welding coupon as the final check

After the geometry passes, assemble a representative joint in the production fixture. Record the actual gap, edge mismatch, assembly force, tack sequence and access. Weld only to an approved procedure and with qualified personnel. This article does not specify welding parameters.

Evaluate the result against engineering and quality-system requirements: assembly stability, fitting effort, visible defects, distortion and the defined non-destructive or destructive tests. If bevel geometry passes but the weld remains unstable, do not alter the laser automatically. Separate the effects of fit-up, fixture, consumables, torch path and welding procedure first.

For robotic welding, also check the seam-tracking range, tool-centre-point access and the accumulated tolerance stack. ART-204 covers the broader preparation of parts for robotic welding; here the robot is only an acceptance condition for the bevel.

Process capability and serial control

Once the measurement system is established, collect a sufficient sample across bars, shifts and batches. Analyse not only the mean but also spread, position trends and outliers. For variable geometry, use a map of permitted values by contour sector.

Define first-part inspection, sampling frequency, recheck after nozzle replacement and a reaction plan. A sample of unknown origin cannot confirm a batch. Every part or batch must link to the material lot, program revision and process baseline.

If statistical capability is required, first confirm that the process is stable, measurement variation is acceptable, and one-sided or two-sided tolerances have been interpreted correctly. A good-looking index does not repair a poor measurement plan.

Typical causes of instability

The first is an incorrect actual centreline caused by ovality or centring error. The second is nozzle or focus drift. The third is insufficient support or regripping. The fourth is a contour outside the kinematic access zone. The fifth is a change in material surface or wall thickness. The sixth is unrecorded manual finishing.

The seventh is measuring only the outside angle. The eighth is using one coupon in a convenient angular position. The ninth is mixing program versions. The tenth is accepting the bevel without mating-fit verification. The eleventh is transferring a supplier demonstration to another model or option.

Define observable evidence for every suspected cause. An offset that repeats by angular position differs from random variation between bars. A root-face trend accompanied by focus drift requires a different response from a mechanical centring error.

What to require from the supplier

Request the exact model, head option, achievable angle range under stated conditions, supported sections, collision rules, CAM licence, sensing, nozzle set, service limits and measurement method. Ask to see the bevel on your material samples and drawing, not only on a demonstration part.

Include several angles, positions and wall thicknesses in factory acceptance; repeat after changeover; measure the inner edge; and perform representative fit-up. Retain the software version, parameter package and raw measurement results. A supplier claim of 3D processing is the beginning of verification, not an acceptance result.

Agree training, application support, calibration, spare nozzles, revalidation and the response to drift. Stable bevel cutting is a system capability, not a single successful part.

When the process is ready

The process is ready when the joint requirement is formalised; the material envelope is known; commanded and measured geometry are linked; the coupon matrix has passed; measurement uncertainty is acceptable; a representative weld is accepted; serial sampling and the reaction plan operate; and the configuration baseline is retained.

Revalidation is needed after a change in tube, supplier, wall thickness, head or nozzle, optics, gas system, jaws or supports, software, joint design, welding process or fixture. If any significant factor changes, the previous pass must be reviewed.

Production traveller for the bevel

In serial production, every batch should link to the material lot, drawing revision, joint code, NC program, machine configuration and approved recipe. The traveller also records the first-part result, operator, time, nozzle and optics state, and disposition of deviations. The data need not all be copied manually; some may be captured automatically, but their relationships must be verified.

Define when the complete profile is measured and when a shop-floor gauge is sufficient. Full metrology may be required for the first part after a changeover, while a control gauge may support periodic sampling. Frequency is determined by process capability and the consequence of failure.

If deburring or cleaning occurs before welding, the traveller must show the condition in which the bevel was measured. Otherwise different shifts may treat geometry before and after a manual operation as the same characteristic.

Diagnosing variation around the circumference

When measured angle changes with angular position, plot it on a polar map. A consistent sinusoidal pattern may indicate centring or a geometric relationship; a local drop near the seam may indicate interaction between material and process; deviation near the chuck may indicate access or support conditions. These are diagnostic patterns, not automatic proof of cause.

Repeat the coupon after changing material orientation if this is safe and included in the plan. If the defect rotates with the material sector, investigate the tube. If it stays in the machine sector, inspect kinematics, optics and setup. Change one factor at a time.

For a variable-angle spatial contour, compare actual results with the nominal map rather than one target. Mark transition zones separately because axis acceleration and the changing effective beam path can make them more sensitive.

Economic criterion for stability

A bevel creates value not when it removes one operation on paper, but when it shortens the route of an accepted part. Count the added cycle, application setup, measurement, consumables, possible lower speed, rework and support. Compare these with mechanical edge preparation and its handling operations.

Include easier fit-up, less manual grinding, stable robotic welding and material flow. Do not attribute savings to the laser before a pilot confirms them. Low, base and high scenarios using actual batch data are more reliable than an advertised percentage.

The release decision should contain both a technical pass and an economic condition. If the bevel is stable but required once a month with a long changeover, another route may remain better. That is not a technical failure; it is correct process selection.

Safe boundaries

This article is not a welding procedure, qualification record or table of cutting parameters. Cutting and welding are performed under manufacturer documentation, a risk assessment and engineering-approved procedures.

Need to verify the bevel-preparation process?