Windscreen adhesive: the specification guide

A windscreen adhesive has one job that matters: keep the glass attached when the car is upside down. Everything else — cure speed, gun time, primer, price — is a constraint on doing that job profitably.

This guide is the map for the windscreen cluster. It states what the bead is asked to do, how the three catalogue grades differ, and where each one stops being the right answer.

The bead is a structural part, not a seal

A bonded windscreen contributes to body stiffness and keeps occupants inside the cabin in a rollover. That is why vehicle makers specify the adhesive, the bead profile and the cure time, and why substituting a “similar” sealant is a change to a safety-critical assembly rather than a maintenance choice.

Two consequences follow, and they explain most of the confusion in this market:

1. Strength alone does not describe the part. A bead in a moving body is loaded in shear and peel, so elongation and modulus decide whether it survives; tensile strength is the headline, not the mechanism.
2. The bead profile is part of the design. The same adhesive applied at half the specified height is a different structure.

What the three catalogue grades actually differ on

Property	KA518	KA520	KA530F	Why it matters
Tensile strength	3.5 MPa	4.5 MPa	6.5 MPa	Headline load capacity of the cured bead
Elongation at break	320%	450%	450%	How much body flex the bead absorbs before tearing
Skin time	30 min	40 min	15 min	How long you have to seat and tool the glass
Cure rate	3.5 mm/day	3.2 mm/day	4.0 mm/day	Time to handling strength through the bead depth
Shore A	65	60	58	Hardness; predicts how shock transfers into the glass edge

The pattern is worth stating plainly: the strongest grade is also the most flexible and the fastest. Higher crosslink density raises tensile and cure speed; the formulation compensates with plasticiser to keep elongation, which lowers hardness. A buyer who assumes “stronger means stiffer” will pick the wrong grade for a bus.

Primerless is a claim about chemistry, not about preparation

Second- and third-generation PU adhesives bond glass, paint and metal without a primer coat. The catalogue states this for all three grades and it is a genuine process saving: one less chemical, one less flash-off window, one less step to get wrong.

It does not remove contamination. Silicone residue from a previous repair, mould release on trim, or grease from a hand on the frit defeats any adhesive, primerless or not. The failure looks identical to a product failure and is not one.

The practical sequence that works: clean with activator, let it flash off, do not touch the frit, seat the glass within the skin time.

The cure window is a factory of its own

Cure rate is quoted at a reference condition — typically 23 °C and 50% relative humidity. Below about 15 °C the reaction slows sharply, and below 5 °C it effectively stops. A cold glass pulls the bead temperature down for the first hour after installation, regardless of the workshop thermostat.

This is the single most common reason a “fast” adhesive behaves like a slow one in winter. The fix is conditioning the glass and cartridges, not buying a different product.

Where each grade belongs

• KA518 — passenger cars, general replacement, a bay that can hold the vehicle for the normal cure window.
• KA520 — buses, coaches and special vehicles: larger glass, higher wind load, more body flex, and a slower installation where a 40-minute skin time is an advantage rather than a cost.
• KA530F — high-throughput bays and cold-weather work where two-hour positioning is the commercial requirement.

When this is not the right choice

• The vehicle maker specifies a two-part system. One-component moisture-curing adhesives are not interchangeable with two-part structural adhesives; follow the OEM sheet.
• The substrate is not glass, paint or metal. Polycarbonate, acrylic and some coated plastics need a different adhesion strategy; test a coupon.
• The joint is a seal, not a bond. A leaking bulkhead seam is a seam-sealer job, not a windscreen adhesive job.

How the specification is actually checked

A windscreen specification has four parts, and they are checked in a different order on site than on paper.

Item	Where it comes from	How it is checked
Adhesive chemistry and grade	Vehicle maker’s sheet	Batch label and certificate
Bead profile (width × height)	Vehicle maker’s drawing	Measure a cut section after removal
Cure time to drive-away	Vehicle maker’s sheet, at a stated temperature	Glass temperature at installation, recorded
Surface preparation	Adhesive supplier’s method	Activator use and frit condition

The gap between paper and practice is almost always in the last two rows. A bay can buy the right cartridge and still fail the job because the glass was cold or the frit was touched.

A failure taxonomy that holds up

• Interface failure at the glass — preparation, not product.
• Interface failure at the frame — paint or primer problem, not product.
• Cohesive failure through the bead — the only pattern that indicts the adhesive or its cure.
• Bead geometry failure — wrong section, wrong profile, or a bead interrupted at a corner.

Classify before calling the supplier. It changes the conversation from a complaint to a diagnosis, and it usually identifies a process fix that costs less than a replacement cartridge.

What the bead actually carries

The numbers make the specification easier to read. Take a 15 kg windscreen in a 20 g deceleration — a modest crash pulse, not an extreme one.

force on the glass      15 kg × 20 × 9.81  = 2,940 N
bond area               4.5 m perimeter × 8 mm bead = 36,000 mm²
shear stress            2,940 ÷ 36,000     = 0.08 MPa

Even at 50 g the demand is about 0.2 MPa. KA518 is rated at 3.5 MPa. The margin is not there because engineers are careless; it is there because the number on the data sheet describes a freshly cured bead at 23 °C pulled in a tensile test, and none of those conditions hold in a car that has sat in the sun for six years.

Three things eat the margin:

• Peel, not shear. A bead loaded in peel concentrates stress along one edge of the bond line. The tensile figure says nothing about peel.
• Ageing at the interface. Moisture and salt reach the glass frit edge first. A cohesive strength of 3.5 MPa with a bond line that has lost adhesion is a 0 MPa joint.
• Cold stiffening. PU gets stiffer as it cools. The same displacement produces a higher stress at −20 °C than at +20 °C, and the glass edge is where it lands.

So the useful reading of a windscreen data sheet is not “how strong” but “how much elongation is left after ageing, and what does the supplier say about adhesion to the actual frit”.

Cure is a diffusion problem, not a clock

A one-component PU bead cures from the exposed face inward. Moisture diffuses into the surface, reacts, and the reaction front moves at roughly the quoted rate — 3.5 mm/day for KA518, 4.0 mm/day for KA530F, both measured by HG/T 4363.

bead height 10 mm, rate 3.5 mm/day
day 1   3.5 mm cured   — 35% of section
day 2   7.0 mm         — 70%
day 3  10.5 mm         — full section

That arithmetic explains a behaviour that looks like a product fault: a car delivered after four hours has a bead that is cured only at the surface. It is strong enough to hold the glass in place, which is what the drive-away time is for, and nowhere near full cure. Anyone who measures “cure” by pressing the bead with a thumb is measuring the first two millimetres.

Two consequences for scheduling:

• Drive-away time and full cure are different specifications. The first is set by the vehicle maker and validated by a crash or pressurisation test on a partially cured bead; the second decides when the joint can be evaluated destructively.
• Deep beads cure disproportionately slowly. Doubling the bead height roughly doubles the time to full cure while the volume doubles with it — the surface-to-volume ratio falls, so the last millimetre is the slowest.

Humidity matters less than temperature at the surface and more than temperature in the core. Below about 30% RH the skin forms slowly and the bead stays tacky longer; above 70% RH the reaction runs faster but the risk of CO₂ foaming rises on thick sections. The 23 °C / 50% RH reference exists because both effects are mildest there.

Temperature is the variable that breaks the schedule

A windscreen bay is heated for people, not for chemistry. Glass arrives from a cold store or an outdoor yard at 5 °C, the bead is applied at 18 °C, and the reaction runs at the temperature of the interface, not of the room.

Condition	What happens	Practical response
Glass below 5 °C	Cure effectively stalls; skin time stretches	Condition glass and cartridges at 18–25 °C overnight
Bay at 10 °C, glass at 20 °C	Cure proceeds; the bead skins while the glass is still being positioned	Work within the skin time, not the room clock
Summer, glass at 45 °C	Skin time collapses; tooling tears the surface	Shade the glass; do not cool it with water
High humidity, thick bead	Faster surface cure, CO₂ foaming risk in the core	Keep bead section within the specified range

The counter-intuitive case is the cold workshop with warm glass: everything looks normal, the bead skins on schedule, and the joint is fine. The dangerous case is warm room, cold glass — the bead skins at the surface while the interface is still at 5 °C, and the first drive-away load finds a weak bond line that no one will inspect for two years.

The rework loop nobody budgets for

Modern vehicles carry a camera behind the windscreen. Replacing the glass moves the optical reference, so the forward-facing camera needs recalibration — static (targets in a workshop) or dynamic (a drive cycle). That work happens after the bead has reached handling strength, which means the vehicle occupies the bay longer than the adhesive cure time alone suggests.

For a bay planning throughput, the sequence is:

1. Remove and clean (30–45 min).
2. Apply and seat (15–20 min).
3. Cure to drive-away (1–3 h depending on grade and temperature).
4. Recalibrate ADAS (20–60 min, static; longer if a road test is required).
5. Hand over.

The adhesive choice affects step 3 only. A fast grade that saves 90 minutes at step 3 is worth nothing if the bay has no calibration equipment and sends the car away, because the customer returns and the second visit costs more than the time saved.

Five questions before you approve a batch

The data sheet answers the easy questions. These are the ones that separate a supplier who controls the process from one who reprints a catalogue:

1. What was the measured skin time and cure rate on the last three batches, and by which method? A method number (GB/T 13477.5 for skin time, HG/T 4363 for cure rate) makes the answer checkable.
2. Which substrates does the primerless claim cover, and was the glass frit among them? Primerless on steel is not primerless on a coated frit.
3. What is the shelf life at 30 °C, not at 20 °C? A warehouse in July is not a laboratory.
4. Does the batch record travel with the pallet? Without it, a failure two years later cannot be attributed.
5. What bead section was the cure rate measured on? A rate quoted for a 3 mm film says nothing about a 12 mm bead.

None of these is a trick question. A supplier who answers all five in one email is telling you how the next dispute will be handled.

The practical position

For a replacement bay, buy one fast grade and one general grade, not three mid-range ones. The fast grade covers winter and throughput; the general grade covers everything where the car can stay. Keep the activator with the adhesive as one system, because the failure mode you will actually meet is contamination, not weak adhesive. And write the batch number on the job card: when a bond fails two years later, the batch record is the only way to separate a product problem from a preparation problem.

Standards referenced

• UN Regulation No. 43 — approval of safety glazing materials and their installation on vehicles. The reason a windscreen is a specified assembly rather than a generic glass panel: the regulation text.
• ISO 4587:2003 — tensile lap-shear strength of rigid-to-rigid adhesives, the test behind most structural bond claims: iso.org.
• GB/T 13477.5 — skin time, the method behind the 15/30/40-minute figures in this guide. Cited by number; the standard is published by SAC and not freely linkable.
• HG/T 4363 — cure rate, the method behind the mm/day figures. Also cited by number.

The catalogue values on the model pages carry these method references; a number without one cannot be compared.