A window or a door does not have one seal. It has five. They sit in different places, they are made by different people, they fail in different ways, and only some of them are covered by anything at all. Almost every claim email we receive opens with the same three words, seal failure, and almost none of them say where the water or the fog actually is. That missing detail is what decides who pays.
| # | Where it is | What it separates | What failure looks like | Who normally owns it |
|---|---|---|---|---|
| 1 | Glass edge seal (insulating glass unit) | The two panes, at the spacer bar | Fog, droplets or a milky film between the panes that does not clear | The glass fabricator, via the window supplier |
| 2 | Glazing rebate seal | Glass to sash | Water sitting in the rebate, stained weep slots, draught at the glass edge | The window manufacturer |
| 3 | Opening joint seal | Sash to frame | Whistling in wind, water on the sill directly below the opening gap | The window manufacturer |
| 4 | Corner and frame joint seal | Frame member to frame member | Water appearing at a frame corner, sometimes with no rain visible on the glass | The window manufacturer |
| 5 | Perimeter seal | Frame to wall | Damp plaster, staining on the wall beside or below the frame, drips following the reveal | The installer and the building envelope. Not the product |
Four of the five live inside the product. The fifth lives in your wall. The first one is made by a glass fabricator rather than by the window factory. That is why a claim that starts with the symptom instead of the location spends its first fortnight going back and forth, and why the photographs you take in the first hour are worth more than the email you write in the first week.
Not every opening has a weather seal that can fail. If the unit in the complaint is a commercial kitchen double swing door, it is a traffic door: it is built to be pushed through from either direction, frequently with hands or a trolley full, and it is not built to be weathertight or draught-tight. There is no gasket line for a claim to attach to, and the right answer is not a warranty decision at all. The same is true of louvred and slatted closures. Before anyone files anything, establish that the product was ever supposed to seal.
When a glass professional says seal failure, they nearly always mean the edge seal of the insulating glass unit. Everything else in the table above is normally called a leak, or a draught. This is the failure that produces fog between the panes, and it is the one worth understanding properly, because its anatomy explains most of the arguments that follow.
Most insulating glass units made today are dual sealed, which means two different materials are doing two different jobs at the same edge.
Published work from the Lawrence Berkeley National Laboratory puts dimensions on this: the primary PIB seal is typically applied at roughly 0.2 to 0.6 mm, the secondary seal at about 4 mm, and a minimum of about 3 mm of secondary sealant is needed to cover the primary seal and shield it from contact with moisture.
The sentence worth remembering from that source is this: PIB adhesion to glass and spacer is not resistant to continuous water exposure. When butyl gets damp, it loses its ability to adhere. The vapour barrier is therefore also the component most easily damaged by standing water, and one of the structural seal's jobs is to keep water away from it. The same body of laboratory literature credits the secondary seal with the largest role in how long a unit lasts.
That has a consequence nobody writes on a drawing. Water allowed to sit against the bottom edge of a unit is not a cosmetic problem to be cleaned later. It is working on the one material that cannot be replaced without replacing the whole unit. This is why the drainage section further down is not a maintenance footnote, and why a blocked weep slot belongs in the same conversation as a warranty claim.
The spacer is filled with a desiccant, most often a molecular sieve or a blend of silica gel with molecular sieve. It adsorbs the moisture trapped in the cavity at the moment of manufacture, and then whatever vapour works its way in afterwards. Its capacity is finite. Once it is spent, vapour that enters condenses on the inner glass surface and you see fog.
So the right mental model is not a wall that either holds or breaks. It is a budget that is spent at different rates. The laboratory literature lists what spends it: temperature fluctuation, solar radiation, atmospheric pressure and wind pressure cycling, oxygen and ozone, aggressive atmospheric gases, cleaning chemicals and surfactants, working loads, and physical attack. A unit on a hot, wet, wind-exposed west elevation spends its budget faster than the identical unit on a sheltered elevation of the same building.
Two failures look like seal failure in a photograph and are not the same thing at all.
Nobody argues about whether a unit is fogged. They argue about whether it was always going to fog, and that question is answered by a test report rather than by an opinion. The useful part is knowing which test answers which question, because suppliers routinely offer the wrong one.
| What you want to know | The test | Standard family |
|---|---|---|
| Is the cavity still dry? | Dew point, or long-term moisture penetration | ASTM E546, EN 1279 part 2, GB/T 11944 |
| Is the insulating gas still there, and how much of it? | Gas concentration and gas leakage rate | EN 1279 part 3, ISO 20492 part 3 |
| Is the edge seal still sound as a material? | Physical attributes of the edge seal | EN 1279 part 4, ISO 20492 part 4 |
| Is the haze water, or a chemical deposit? | Chemical fogging | ISO 20492 part 2, ASTM E2189 |
| Will the next batch be the same? | Factory production control and periodic tests | EN 1279 part 6 |
| How does the whole unit perform and get evaluated? | Overall performance and evaluation | ASTM E2190, ASTM E2188, ISO 20492 part 1 |
Three rules make this table usable in a negotiation. First, a result without a standard named on it is not a result. The word tested tells you nothing, because everyone tests something. Second, ask which specimen was tested. A dew point result obtained on a small, easy unit says very little about the large, heavy, dark-tinted pane actually going into your west elevation, and the edge seal works hardest exactly where the glass is biggest and the frame is hottest. Third, a certificate attaches to a plant rather than to a brand. Published guidance on qualifying a glass supplier makes the point directly: a certification held at one facility does not automatically cover a manufacturer's entire global network, and the public listings maintained by bodies such as the Insulating Glass Certification Council and the Safety Glazing Certification Council should be checked against the plant that will actually make your glass.
This is the most common misunderstanding we see in specifications, and it costs people claims they did not need to lose.
Guardian Glass, in its own technical literature on insulating glass units, states plainly that it is natural for the fill gas to diffuse through the edge seal over time, that an accepted diffusion rate is one percent per year, and that even at that rate the remaining gas continues to provide insulating benefit. Diffusion is not the same event as a failed seal.
So the question in a claim is not whether the gas is still at its fill level. It is whether there is visible moisture or deposit in the cavity. Those are two different tests answering two different questions. EN 1279 part 3 covers the long-term test method for gas leakage rate and gas concentration tolerances. ASTM E2190 covers insulating glass unit performance and evaluation. ASTM E546 covers the dew point test method.
The practical instruction is dull and it works: if you care about gas concentration, write the target and the standard it is measured under into the order. A unit can be perfectly sealed and still fail a number that nobody ever wrote down.
| Where you see it | What it actually means | Is it a claim |
|---|---|---|
| On the room-side surface of the glass | Indoor humidity meeting a cold surface. A heating, ventilation and occupancy question | No. This is the single most common thing mistaken for a seal failure |
| On the exterior surface of the glass | The outer pane is cold because the unit is insulating well. Published glass warranty wording commonly lists exterior condensation as a normal occurrence | No |
| Between the panes, inside the cavity | Moisture or a deposit in a space that was sealed dry at the factory and cannot be wiped | Yes. This is the one that is normally a seal claim |
The test that settles it in thirty seconds: if you can wipe it, it is not in the cavity.
Where it has to be settled formally, laboratories use a dew point test, in which the outer pane is progressively cooled while the cavity is watched for fog. EN 1279 part 2 covers the long-term test method and requirements for moisture penetration, while GB/T 11944 is the equivalent Chinese standard and AS/NZS 4666 and JIS R 3209 exist in their own markets. Ask which one your report is written against, because a pass under one standard does not translate into a pass under another, and a converted result is not a tested result.
The fifth seal in the table is made of your sealant, your backer rod, your flashing or membrane, your sill pan, your anchorage, and the drainage path that has to stay open to the outside. Published commercial glazing warranty guidance splits coverage into three separate layers, glass, framing and installer workmanship, and is explicit that building movement beyond design tolerance is excluded.
Why it matters more than any clause: the perimeter is the only one of the five seals made on your site, by your trades, out of your materials, after our crate has been opened. We can supply profiles, gaskets, drawings and test evidence. We cannot sign off a joint we did not make, and we cannot act as the applicant for approvals in your jurisdiction.
If water shows up on the plaster beside or below the frame, the first question is not whether the window is faulty. It is whether the perimeter was built as a drained assembly or sealed up as a box. Sealing the outer perimeter all the way round is one of the best documented ways to trap water inside the assembly until it finds its way indoors, and it is usually done by whoever is trying to fix a leak rather than by whoever caused it.
The sash-to-frame seal behaves completely differently depending on how the sash moves, and this is where most arguments about sliding units begin.
That is why our sliding range leans on drainage hardware rather than on sealing adjectives. SL126/182 offers a 45 mm high-ｄｒｏｐ bottom track and an optional 85 mm high water barrier, with low track 8 mm and high track 43 mm options. SL155 carries passive check valve drainage. SL126/182(75) carries a butyl rubber waterproof gasket at the frame-sash joints.
TS130 exists because a plain sliding sash cannot compress a gasket. The side-press mechanism moves the sash out of the sliding plane so that the seal is compressed rather than brushed. Its published wording, casement-level sealing performance, is a comparison against a sliding pile seal. It is not a claim to be a casement window, and it should never be read as one in a specification.
The honest consequence for a claim: a sliding sash that feels draughtier than a casement on a windy day is often reporting the operating principle rather than a defect. The defect question is whether the contact seal is present, intact and correctly adjusted. That is a different question with a different answer, and usually a far cheaper fix.
The bottom track of a sliding unit is a water management component. It is meant to collect whatever gets past the outer line and take it back outside through weep slots or a check valve.
Standing water in the outer track after rain is the system doing its job. Water crossing the inner upstand is the failure. Those two sentences resolve a large share of sliding door complaints without a single part being replaced.
The maintenance item that quietly voids claims is the outlet. Weep slots and drainage paths get blocked by mortar, render, paint, silicone and site debris, and published glass warranty exclusions routinely name concrete and stucco overspray, welding spatter and pressure washing among the things that are never covered. All three are site operations, not product conditions.
Every drainage feature we publish has the same dependency: the outlet must stay open and must not be sealed over by the finishing trade. Put that sentence in the subcontractor's scope rather than in the window's specification, because the person who blocks a weep slot is almost never the person who ordered the window.
| What happens | Which seal it attacks | Who usually does it |
|---|---|---|
| Pressure washing the frame or the glass | Perimeter and glazing rebate, and it drives water against the glass edge seal | Cleaning contractor |
| Razor blades, abrasive pads or ammonia on coated glass | Glass surface and coating | Cleaning contractor |
| Aftermarket film or tint applied to the glass | Glass edge seal, through added thermal stress | Owner or tenant |
| Mortar, render and concrete overspray | Weep slots, drainage path and coating | Site trades working nearby |
| Welding spatter from adjacent works | Glass and coating, permanently | Adjacent trades |
| Building movement or settlement beyond design tolerance | Perimeter, and frame geometry | Structure and foundation, not the window |
| Cleaning chemicals and surfactants used routinely | Edge seal chemistry over time | Maintenance team |
| Any repair attempted before inspection | Whichever seal the repair touched, and the evidence | The claimant, meaning well |
The clause most often broken is the one that sounds obvious: do not attempt repair before inspection. The natural response to a leak is to seal it. Once it is sealed, nobody can tell which of the five seals was leaking, and the claim has lost the only evidence that mattered.
| Symptom | Which seal | Usually a claim | First action | What to photograph |
|---|---|---|---|---|
| Fog or milky film between panes that does not clear | Glass edge seal | Usually yes, subject to term and exclusions | Report and photograph. Do not drill the unit and do not attempt to dry the cavity | The whole unit including one corner, plus a close-up of the spacer line, in daylight |
| Frost or droplets inside the cavity in cold weather | Glass edge seal | Yes, same conditions | Same, and note the outdoor temperature and the time of day | The cavity at an angle to the light, plus the same unit from outside |
| Water on the sill directly below the opening gap | Opening joint or perimeter | Depends entirely on where it enters | Dry the area, then wet the wall above only, not the glass. Then wet the glass only. The order tells you which seal it is | A wide shot showing frame, sill and wall together |
| Water appearing at a frame corner | Corner joint | Usually yes | Photograph before touching anything | The corner from inside and outside, with a coin or tape measure for scale |
| Damp plaster beside or below the frame | Perimeter | Almost never a product claim | Check the external perimeter bead and the weep outlets first | The external reveal and the weep slots, plus the interior damage |
| Draught at the glass edge, no water | Glazing rebate, or lock adjustment | Often an adjustment, not a defect | Check lock point adjustment before ordering any part | The closed sash from inside, plus a close-up of the lock points |
| Draught through a sliding sash in wind | Contact seal, operating principle | Usually no | Confirm the contact seal is present, intact and correctly adjusted | The interlock and the track, taken from the end of the unit |
| Staining below the weep slots | Drainage path | No | Clear the outlet and re-check after the next rain | The weep slot close up, before and after clearing |
And now the item that causes more delay than anything else: unit identification. The order line has to let us find the unit in a production record. A line that reads only a 36 x 78 entry door gives us the call-out, not the unit. The purchase order needs the millimetre frame size, the configuration, the glass build-up, the colour batch and the opening reference. Without those we can quote you a replacement but we cannot reproduce the original, and for a seal failure the reproduction is the whole point.
| What we publish | What it actually measures | What you must still ask for |
|---|---|---|
| Triple sealing on C113 and C116, with seamless welded corner sealing on C116 | A sash-to-frame sealing arrangement | The tested air permeability and watertightness class, and the standard it was tested under |
| Dual-Matrix Eight-Layer Sealing on AW90 | The sealing stack inside a tilt-and-turn sash | The classification report. A layer count is not a test result |
| Four-sided continuous sealing strip plus upper-rail frame sealant on TS130, and side pressure rubber strip | The sash-to-frame interface of a side-press sash | Whether the quoted sealing figure applies to the side-press configuration or to a fixed panel |
| Butyl rubber waterproof gasket at frame-sash joints on SL126/182(75), and continuous four-sided sealing | A joint sealing material at the frame-sash junction | The glass build-up and the insulating glass edge seal specification. That is a separate component with its own maker |
| 240-hour salt spray corrosion resistance, AW90 | Coating and substrate corrosion resistance | Nothing from this number. It is not a seal rating and it does not transfer to another series |
| 5,000 N shear resistance, AW90 | Thermal break strip to aluminium composite bond | Nothing from this number either. It is a composite bond figure, not a seal figure |
| Drainage geometry: 45 mm high-ｄｒｏｐ bottom track, optional 85 mm water barrier, 8 mm and 43 mm track heights, passive check valve | Water management geometry | A watertightness class, and a written statement of which trade keeps the outlets clear |
| Injection-moulded structural adhesive sealing on C120, glued corner brackets on C113 | A corner construction method | Confirmation that the corner detail on your order is the one shown in the approved drawing |
The pattern worth noticing: the two hard numbers we publish in this area measure corrosion resistance and a composite bond. Neither is a seal rating, and neither transfers from AW90 to any other series. Everything else in that table is a description of construction. A construction description is useful, it is also what every competitor prints, and it is not a test result. Only a test result decides a claim.
First, if your schedule calls for folding outdoor doors, or for entry doors in the sizes usually catalogued separately, our folding door and entry door categories are currently empty. We will tell you that rather than offer a sliding unit as a substitute, because a substituted product is exactly where seal arguments begin.
Second, we can supply material and construction evidence, drawings, sections and test reports. We cannot be the applicant for a local approval, we cannot certify a perimeter joint made by your trades, and we will not put a number in writing that we cannot reproduce from a test report. If a supplier offers you any of those three, ask what exactly they are signing.
Ten lines to paste into your next enquiry or purchase order. The right-hand column is the part that saves money.
| Line | What to write | A useless answer looks like |
|---|---|---|
| 1 | Glass build-up in millimetres, including which surface carries the coating | Double glazed |
| 2 | Spacer type, and whether it is a warm-edge spacer | Standard spacer |
| 3 | Primary and secondary sealant types, named | High quality sealant |
| 4 | Desiccant type | Contains desiccant |
| 5 | Gas fill type, target concentration, and the standard the concentration is measured under | Argon filled |
| 6 | Dew point or moisture penetration test standard, and the result | Fully tested |
| 7 | Confirmation that the coating is deleted at the edge before sealing | Low-E as standard |
| 8 | Air permeability and watertightness class, with standard, and whether the test specimen was the opening configuration | Meets international standards |
| 9 | Who owns the perimeter seal, the weep outlets and the drainage path, named by trade | Installer to seal as required |
| 10 | Prohibited cleaning methods and chemicals, listed explicitly | Clean regularly |
It is the symptom that almost every glass warranty recognises, because the cavity was sealed dry at the factory and cannot be contaminated by anything on site. But coverage still depends on the term, on the exclusions, and on whether anyone has modified the glass. Report it as soon as you see it, and do not drill the unit. Drilling turns a straightforward unit replacement into an argument.
Not on its own. Published insulating glass literature from a major glass manufacturer treats a slow diffusion of fill gas through the edge seal as normal, at an accepted rate of around one percent a year, with the remaining gas still delivering benefit. The trigger for a claim is visible moisture or deposit inside the cavity. If gas concentration genuinely matters to your project, specify it as a number with a test standard attached, at order stage.
Almost certainly not. If you can wipe it, it is on the room-side surface, and it is a humidity and ventilation question rather than a product question. Condensation on the outside surface is, if anything, a sign the unit is insulating well. Only moisture inside the cavity is a seal failure, and that is the one you cannot wipe.
No, and no manufacturer's product warranty does. That joint is made on site, out of site materials, by a different trade, and published warranty structures treat installer workmanship as a separate layer. We can supply drawings, profiles and the test evidence for what we made. We cannot warrant a joint we did not make.
No. The edge seal is manufactured into the unit under factory conditions, and once it has failed the insulating glass unit is replaced rather than repaired. That is the practical reason the unit identification data matters so much: we are not quoting you a generic piece of glass, we are reproducing a specific one.
Settle it before you need it, not after. Cross-border replacement parts are commonly shipped with duties unpaid on the buyer's side, so the replacement can arrive with a bill attached. Ask for the shipping term, the customs description and who carries the import charges, and get all three in writing when the order is placed. Also ask whether an advance replacement is available, so that the site is not waiting on the old unit being returned first.
