Cracked 2024-T3 bracket: weld, braze, or replace?

Reject the weld.

• 2024 is a copper-bearing (about 4.4 % Cu) aluminium alloy and is not regarded as fusion weldable. The copper-rich grain-boundary constituents melt at a lower temperature than the bulk, giving hot-short cracking - weld metal and heat-affected zone cracks propagate as the weld solidifies and contracts.
• The T3 temper is produced by cold work plus natural ageing. Welding destroys that condition locally, so the heat-affected zone reverts to near-annealed strength - roughly a third to a half of the parent metal - leaving a soft notch exactly where the load path was interrupted by the crack.
• If the material is Alclad, the pure-aluminium cladding melts first and the copper dilution changes the weld chemistry unpredictably; the cladding's corrosion protection is also destroyed either side of the weld.
• Weld repairs to structural brackets are only acceptable against approved data; there is none here, and a hand-finished weld with no post-weld heat treatment cannot be shown to restore the drawing's allowable stresses.
• Residual stress plus a drilled-hole stress raiser is an ideal fatigue initiation site - the crack will return.

Brazing: worse than welding in practice. Zinc-based aluminium brazing needs an aggressive fluoride/chloride flux; if any residue is trapped in the crack or under the lap it will cause severe corrosion in service. Brazing temperature (about 380–450 °C) is still high enough to overage the 2024 and destroy the T3 condition locally, while giving none of the strength of a weld. No strength credit can be taken, and there is no approved data. Reject.

6061-T6 replacement, welded: 6061 is a genuinely weldable alloy, so this is the only proposal with any metallurgical merit - but it is not a like-for-like swap:

• 6061-T6 parent metal is roughly 290–310 MPa UTS; the welded HAZ in the as-welded condition is about 165–170 MPa, i.e. around half. A weld in a corner bracket with a bolt hole may be overstressed unless the design is checked.
• Full properties require post-weld solution treatment and artificial ageing (or the part redesigned thicker / with the weld away from the load path), and the finish schedule (anodise or alodine plus primer) must be applied after welding.
• Substituting material and process changes the approved design and needs approved data/engineering disposition, not a shop decision.

Correct action: fabricate a new bracket to the drawing, in the specified 2024-T3, rather than repair the cracked one. The bracket is simple, not primary structure, the material and drawing are available, and a new part restores the original design with no unapproved process.

If a repair is genuinely preferred, the acceptable route is a riveted repair (crack stops plus a doubler) or a weld on an approved alternate alloy - either way against approved data, with:

• crack-stop holes at the crack tips, deburred;
• doubler thickness and rivet size derived from the scheme, pitch 4D–6D, edge distance at least 2D, rivet diameter between 2t and 3t of the thinner sheet;
• corrosion protection between dissimilar parts (e.g. cadmium-plated or primed fasteners, wet assembly with sealant) and full finish restoration;
• the load path checked - a doubler adds eccentricity and secondary bending.

Before release: confirm the material certificate for the replacement stock (2024-T3, not a substitute), confirm the drawing revision, record the fabrication and any repair against approved data in the tech log, apply the specified finish, and have the task certified by the appropriately authorised person. Do not let the three-day schedule drive an unapproved weld on a heat-treated alloy - that is exactly the decision the regulation exists to prevent.