Glazing & Curtain Wall
QA inspection checklist

Frame Installation

1. 1.

   Framing system installed to approved shop drawings
   What to look forSections, profiles, and connections match shop drawings (latest revision). System manufacturer/series identified on profile labels. CWCT test data and Declaration of Performance available for system.
   Acceptable whenPer CWCT Standard / BS EN 13830 (curtain wall standard). Approved shop drawings (latest revision). System tested per CWCT TN 76 / BS EN 13830 — air permeability, water tightness, wind resistance.
   StandardBS EN 13830:2015+A1:2020 — Curtain walling product standard. CWCT Standard for Systemised Building Envelopes. CWCT Technical Notes.
   Common defectsProfile from another manufacturer's system mixed in — no warranty/testOld shop drawing revision used — clashes with structureDeclaration of Performance not available — non-CE/UKCA compliantGeneric "off-the-shelf" sections without project-specific testing
2. 2.

   Mullions plumb and transoms level

   Record: Deviation (mm)
   What to look forPlumb checked top-to-bottom of building per mullion. Transoms level using rotary laser. Alignment between adjacent units consistent. No racking or distortion of frames.
   Acceptable whenPer CWCT Standard cl.3.5: mullion plumb ±5 mm in storey height, ±15 mm overall. Transom level ±2 mm in 3 m, ±5 mm overall in line. Mullion-to-mullion alignment ±2 mm.
   StandardCWCT Standard for Systemised Building Envelopes Section 3. BS EN 13830:2015+A1.
   Common defectsMullion lean accumulating up building — base setting-out offTransoms not level — water flow direction reversed in drainageAdjacent mullions out of line — visible jog in finished façadeFrames distorted under self-weight — brackets undersized
3. 3.

   Brackets and fixings installed to structural design
   What to look forBracket type and orientation matches structural detail. Fixings into structure at correct depth and torque. Slotted holes used where movement joints designed. Stainless steel grade per CWCT (typically 316 for coastal, 304 inland).
   Acceptable whenBrackets per design — typically aluminium or galvanised steel for vertical, stainless for façade-fixed. Anchor pull-out tested in accordance with ETAG 001 / BS EN 1992-4. Slotted holes for thermal/structural movement.
   StandardBS EN 1992-4:2018 — Design of fastenings. ETAG 001 — Metal anchors. CWCT TN 31 — Curtain wall fixings.
   Common defectsWrong stainless grade — 304 specified in coastal, surface rust within 12 monthsSlotted holes packed with washers — movement preventedAnchor torqued before resin cured — pull-out failureFixings not pull-out tested on siteBracket spacing exceeds design — wind load fails
4. 4.

   Thermal breaks intact and continuous
   What to look forPolyamide or polyurethane thermal break visible at every aluminium junction between inner and outer face. No bridges (metal cleats spanning the break). Continuous around frame perimeter and at corners.
   Acceptable whenThermal break per system test data — typical Uf ≤ 1.6 W/m²K for new build. Polyamide ≥ 16 mm depth typical. No metal-to-metal bridges (e.g. fixing brackets, cleats). U-value (Ug, Uf, Uw) per design.
   StandardBS EN ISO 10077-1:2017 — Window/door thermal performance. BS EN 14351-1 — Windows external doors. CWCT TN 32.
   Common defectsMetal screw or cleat bridging thermal break — cold spot, condensationPolyamide insert missing in adapter sectionSteel reinforcement bonded directly to inner aluminium — thermal bridgeDrilling for ironmongery cuts through thermal break
5. 5.

   Drainage channels clear and weep holes open
   What to look forMullion and transom drainage channels free of debris, sealant, or insulation. Weep holes/slots present at base of each light, free-flowing. Drainage falls correctly to outside. Pressure equalisation slot also clear.
   Acceptable whenPressure-equalised drained system per BS EN 12208 (water tightness) / CWCT Standard. Weep slots at base of each pane, typically 2 per metre minimum. Drainage path tested with water during commissioning.
   StandardBS EN 12208:2000 — Watertightness of windows and doors. BS EN 13830. CWCT Standard.
   Common defectsSealant blocks weep holes — water trapped, eventually leaks insideDrainage channels filled with mortar/insulation debrisNo pressure equalisation slot — water sucked in under windWeep holes facing wrong way (upwind) — driven rain entersSingle weep per panel — blocks easily and back-up

Glass & Panels

1. 6.

   Glass units match specification
   What to look forEtched mark on each pane gives manufacturer, date, build-up (e.g. 6/16/4 = 6mm + 16 cavity + 4 mm). Toughened/laminated/heat-soaked stamped. Coatings (Low-E, solar control) on correct face per drawing — typically face #2 or #3.
   Acceptable whenIGU to BS EN 1279 (sealed unit), toughened to BS EN 12150-1, laminated to BS EN 14449. Heat-soaked toughened (HST per BS EN 14179) for safety in overhead/façade. U-value, g-value per design.
   StandardBS EN 1279 series — Insulating glass units. BS EN 12150-1 — Toughened safety glass. BS EN 14179 — Heat soaked toughened. BS EN 14449 — Laminated. BS 6262.
   Common defectsToughened not heat-soaked where overhead — nickel sulphide spontaneous breakage riskCoating on wrong face — solar gain wrong direction, condensationStandard annealed used where toughened/laminated specified — safety glazingBuild-up not as spec — 4-16-4 instead of 6-12-4 affects thermal/acousticEtched mark missing — no traceability
2. 7.

   No chips, scratches, or defects
   What to look forExamined from 3 m distance against neutral background. No bubbles, inclusions, scratches, edge chips. No "dragged" digital print or coating defects. Edge condition (toughened) — no nibbles or shells.
   Acceptable whenPer BS EN 12150-1 / GGF good practice: no scratch > 75 mm long visible at 3 m. Edge chips ≤ 2 mm in width and depth. No bubbles, inclusions, or dirt within IGU. Coating uniform.
   StandardGGF "Quality of Vision through Glass". BS EN 12150-1. BS EN 1279-2.
   Common defectsEdge chips on toughened — risk of spontaneous fractureRoller-wave distortion visible (toughened) — line of distortion under low-angle lightScratch from suction cup damage during installNewton's rings (interference) showing IGU pressure issuesLogo/etched mark in vision area
3. 8.

   Setting blocks correctly positioned
   What to look forSetting blocks at quarter points of pane (or per system spec). Block material correct — nylon or similar (not wood, not metal). Block size matches glass thickness — full glass thickness with edge clearance.
   Acceptable whenPer BS 6262 / GGF: setting blocks at quarter points or per system; minimum 2 blocks per pane. Material EPDM, neoprene, or nylon. Size width ≥ glass thickness, length 50-100 mm typical. Hardness Shore A 80-90.
   StandardBS 6262-6:2018 — Glazing for buildings. GGF Glazing Manual. CWCT TN 35.
   Common defectsSingle setting block at centre — pane rests on edge, breakageWood/timber block — degrades, leaches into IGU primary sealBlock too narrow — supports only inner pane of IGUNo block at all — glass sits on aluminium, breakage at cornersSoft rubber compresses — pane drops, top edge bites frame
4. 9.

   Glazing gaskets seated fully
   What to look forEPDM gaskets seated full depth in groove — no kinks, gaps, or twists. Welded corners (factory) preferred over butted. No daylight visible at corners. Replacement gaskets fit snugly without forcing.
   Acceptable whenGaskets to BS EN 12365-1, hardness, recovery, and durability classes per CWCT. Fully seated to gasket groove depth. Corner welds preferred — butted joints with adhesive sealant accepted only as substitute.
   StandardBS EN 12365-1:2003 — Building hardware — Gasket and weatherstripping. CWCT Standard.
   Common defectsGaskets pulled tight at corners — rubber stretches and shrinks back leaving gapsButted corner joints unsealed — water entry pathWrong gasket profile — wrong-sided pressure plateOld gaskets re-used — set-permanent, no recoveryTwisted gasket — compression seal lost
5. 10.

   Spandrel panels colour matched and aligned
   What to look forSpandrel panel colour matches vision glass back paint or specified RAL. Insulation visible from outside? — secondary coating or shadow box detail. Aligned in plane with vision glass. No visible printed pattern or backing through.
   Acceptable whenSpandrel infill per CWCT TN 28 — opacified IGU, ceramic frit, paint-back, or shadow box. Colour to RAL/NCS spec. Insulation behind ≥ 25 mm thermal break to air void. Visual: indistinguishable from vision glass at 3 m unless specified.
   StandardCWCT TN 28 — Spandrel panels. BS EN 13830. BS EN 1279.
   Common defectsInsulation visible from outside — translucent paint-back, dark visibleHeat build-up in shadow box causing seal failure of IGUColour off RAL by 2-3 shades — visible against adjacent vision glassSpandrel panel proud of vision glass — visible step in façade

Sealing & Weatherproofing

1. 11.

   Perimeter sealant applied continuously
   What to look forSealant joint continuous around frame — no breaks at corners, transoms, or interfaces. Tooled to firm pressure — no air voids. No bubbling or skinning during application. Joints prepared (cleaned, primed where needed).
   Acceptable whenPer CWCT TN 13 / BS EN 15651: sealant joint continuous. Width ≥ 6 mm typical, depth ½ width with backer rod. Cure time per manufacturer. Adhesion test (peel test) on completed joints recommended.
   StandardBS EN 15651-1:2017 — Sealants for façade elements. CWCT TN 13 — Glazing sealants. ASTM C920 (US equivalent).
   Common defectsSealant skipped at internal corners — water entry pathNo primer on aluminium — adhesion failureThree-sided adhesion (silicone bonded to backer rod) — splitsJoint width inadequate — sealant tears at first thermal cycleSealant tooled with wet finger — surface contaminated
2. 12.

   Sealant type matches specification
   What to look forCartridge label matches spec — manufacturer, product (silicone, polyurethane, MS polymer), grade. Structural silicone clearly labelled (typically Dow 995 / Sika SG 500 / Tremco PS). Use-by date in date.
   Acceptable whenSealant per BS EN 15651-1 (façade) / EN 15651-2 (glass). Structural silicone separately tested per ETAG 002. Type per spec — neutral cure silicone for IGU edges, polyurethane for movement joints. Compatibility tested with adjacent materials.
   StandardBS EN 15651-1 to 5. ETAG 002 — Structural sealant glazing. CWCT TN 13.
   Common defectsAcetic-cure silicone (smells like vinegar) on metal — corrosionDomestic silicone used on façade — UV failure within 3-5 yearsOut-of-date sealant — slow cure or no cureWrong colour — visible against curtain wallMixed product brands within same joint — incompatibility risk
3. 13.

   Backer rod installed at correct depth
   What to look forClosed-cell backer rod sized at 25% larger than joint width. Set to depth giving sealant depth ½ joint width. Bond breaker tape used at narrow joints where rod will not fit.
   Acceptable whenBacker rod closed-cell PE per BS 6213. Diameter 25% greater than joint width. Sealant depth ½ × width (e.g. 10 mm wide joint = 5 mm depth). For joints < 10 mm depth, bond breaker tape used.
   StandardBS 6213:2000 — Sealing joints in buildings. CWCT TN 13.
   Common defectsBacker rod punctured during install — outgassing causes bubbles in sealantNo backer rod — sealant fills full depth, three-sided adhesion, splitsOpen-cell rod used — absorbs moisture, sealant fails to cureWrong size — too small, falls into cavity; too large, prevents sealant entering
4. 14.

   Water test completed (if required)
   What to look forHose test or AAMA water test per CWCT Standard. Spray bar 1 m from façade, water at controlled flow rate. Inspector inside watching for leaks. Each section tested in turn. Failed sections rectified and re-tested.
   Acceptable whenPer CWCT TN 41 / AAMA 501.2: spray nozzle 1 m from façade, 2.5 l/min/m at 200 kPa, 5 min per location. No water penetration to interior. Static air/water tests per BS EN 12153/12155 for full-scale tests.
   StandardCWCT TN 41 — Site testing. AAMA 501.2 — Hose test. BS EN 12153/12155 — Air/water test.
   Common defectsHose test pressure too low — does not reproduce wind-driven rainTest bypassed because programme tight — leaks discovered post-handoverFailed section rectified but not re-tested — unknown if fixedInsufficient duration — slow leaks missedTest only at sample locations not full façade

Run this inspection on a phone

The same 14 checks, with the guidance a tap away, photos against each fail, measurements recorded, an NCR raised automatically and a branded PDF at the end. Signed on site. 14 days free.