The leak usually starts above the stain

When an aluminum window shows damp at the top corners or a brown line under the head reveal, the first instinct is to blame the frame seal or the glass. On site, that guess is often wrong. The water that shows up inside usually entered somewhere else in the assembly, then traveled until it found the lowest-pressure exit. An aluminum window head detail that is built as a layered system can keep that water moving outward. A head detail built as a single bead of sealant cannot.

The useful way to think about the head is simple: it is a drainage and control zone, not a cosmetic joint. If the layers above the window are arranged correctly, the exterior can get wet without the interior getting damaged. If the layers are out of order, the assembly has no safe way to shed water, and the leak may not appear until the first major storm, the first winter of repeated wetting, or the first time a sealant joint ages enough to crack.

Sealant is the last line, not the first

A bad habit in many projects is to treat the visible sealant bead as the whole answer. That approach is appealing because it is fast, cheap, and easy to inspect from the ground. It is also the reason so many head leaks repeat after repairs.

Sealant has one job: help the joint resist incidental moisture and air movement. It is not designed to replace:

  • a flashing that sends water back outside
  • a cavity tray that catches what gets through the outer leaf
  • stop ends that keep water from escaping sideways
  • a drained cavity that can actually empty
  • an air seal that blocks warm moist air from entering the cold zone

Sealant is maintenance, flashing is architecture. If those layers are missing, sealant becomes the primary defense. The minute it shrinks, debonds, or is cut by movement, the wall has no backup.

That is why two windows can look identical from the room side while performing completely differently in rain. One has a functioning head junction behind the trim; the other has a bead over a gap and nothing else.

Why the visible failure is rarely the real entry point

Water does not always enter where the stain appears. It often enters above the window head, then runs along a lintel, a membrane lap, or the underside of a flashing until it finds a path inward.

That is especially common in three situations:

  • Cavity masonry: A cavity tray is installed with the wrong fall, missing stop ends, or blocked weeps. Water that should have drained out is trapped above the opening and eventually appears at the top of the frame.
  • Timber frame with cladding: The head flashing is not lapped under the weather-resistive barrier. Water gets behind the flashing and follows the wrap or sheathing until it reaches the window opening.
  • Rainscreen construction: The head flashing stops flush with the cladding face instead of projecting far enough to shed water. Wind-driven rain tracks back into the cavity instead of away from it.

In all three cases, the external bead can look fine. The real defect is higher up, deeper in, or both. That is why chasing a bead replacement alone often wastes time and money. The visible joint is usually only the last place the water was able to escape inward.

The order of installation decides whether the assembly works

A head detail is not just a list of materials. It is a sequence. Get the order wrong and even good products fail.

The reliable sequence is straightforward:

  1. Structural support is installed level and capable of carrying the load above the opening.
  2. The drainage layer or flashing is placed so water can move outward, not inward.
  3. Insulation and air barrier continuity are carried through the head so the cold zone is not exposed.
  4. The window frame is fixed with the correct clearance so movement does not crush the joint.
  5. External weather seals are applied only after the substrates are dry, clean, and correctly lapped.
  6. Internal air seals are closed before the lining hides the junction.

If any one of those steps is reversed, the detail loses redundancy. For example, if the frame is installed before the membrane is properly lapped, the membrane often ends up cut, folded the wrong way, or simply stopped short. If sealant is applied before the substrate dries, adhesion failure becomes a timing issue rather than a material issue. If the frame is packed tight against the lintel, thermal movement and structural deflection can crush the joint and split the seal later.

This is why a good installer does not ask, “Can I make this look sealed?” The better question is, “If the sealant fails next year, does the assembly still shed water?” That question is the real test of a window head detail.

A proper head detail can survive a seal failure

The strongest proof that a head detail is well designed is not that the sealant looks new. It is that the assembly still has a path to drain if the sealant ages out.

A durable head detail usually has these traits:

  • Water is intercepted before it reaches the frame head
  • The flashing or tray has a defined fall to the exterior
  • Stop ends prevent side leakage
  • Weep points allow collected water to exit
  • The frame is not hard-packed against the support above it
  • Interior air sealing is continuous enough to prevent hidden condensation
  • Insulation bridges the head so the cold surface is not exaggerated

When those conditions are present, the sealant becomes what it should have been from the start: a secondary barrier, not a life-support system.

This distinction matters because sealants are maintenance items. They age under UV, move with temperature, and eventually need replacement. A correctly detailed head does not depend on a perfect bead for the building to remain dry. It depends on geometry, laps, and drainage.

The fastest way to diagnose a bad head detail

When an aluminum window keeps leaking at the top, the most useful questions are not cosmetic. They are structural and sequence-based:

  • Can the wall above the opening shed water without touching the frame?
  • Does every upper layer overlap the one below it?
  • Are the stop ends sealed, or can water run off the sides of the tray?
  • Is the frame separated from the lintel by a tolerable gap, or is it crushed into place?
  • Does the head still drain if the external seal is ignored?
  • Can moisture inside the wall escape outward, or is it trapped behind a closed joint?

If the answer to that last question is no, the problem is not a poor sealant bead. The problem is that the detail was never redundant enough to be called a detail at all. It was just a joint waiting for the weather to prove it.