What ICP-002 sets out to do
ICP-002 is an industry code from the AGWA that standardises how the longevity-critical properties of window and door hardware are tested and rated for housing and residential buildings.
Rather than declaring a single "acceptable" product, the code gives everyone in the supply chain — the hardware maker, the window fabricator, the specifier and the end consumer — a shared way to compare like with like. It does that by converting two real-world failure modes into graded scales: a corrosion resistance rating and a cyclic durability rating. A latch rated CR6 / D3 can then be weighed against one rated CR8 / D6 on the same terms.
The code is deliberately narrow. It concentrates on the two things that most determine how long hardware keeps working: how a finish and its substrate hold up against salt-driven corrosion, and how many operating cycles a mechanism survives while staying smooth to use. It does not assess weather-tightness or wind-load resistance of the finished opening — those belong to the window and door system as a whole and are handled by testing to AS 2047. The committee has flagged water-penetration and structural performance as candidates for future revisions.
Comparison, not substitution
ICP-002 ratings let you rank one item of hardware against another. They are not a licence to swap hardware inside a given window or door system — whether a substituted item keeps the system compliant remains the responsibility of the window or door manufacturer. Routine preventative maintenance to the maker's instructions is assumed throughout, and is essential for hardware to reach its intended service life.
What the code covers
ICP-002 applies to hardware fitted to external residential windows and doors as defined in the National Construction Code. Throughout the document the word "sash" is used to mean both windows and doors. Hardware governed by the swimming-pool safety-barrier standard (AS 1926.1) sits outside its reach.
Opening types in scope
Hardware for any of the following sash actions:
Hardware items in scope
The mechanisms and fittings that make a sash move, hold and lock:
Not every item faces every test. A segregation table in the code maps each hardware type to the tests it must sit — corrosion, cyclic durability, ease of operation, or a combination. Balances, for example, are checked for corrosion and cyclic durability but are exempt from the ease-of-operation force limits, since their job is to counterweight rather than to be operated by hand.
Two independent streams, one certificate
Running a salt-spray attack and a hundred-thousand-cycle workout on the same physical sample would punish it far harder than any real installation ever would. So the code splits testing across two parallel streams, each using a separate but identical production sample:
Stream A — Corrosion
One sample is mounted as it would be installed and exposed to a neutral salt-spray environment for a duration set by the target corrosion grade, then checked for finish and function.
Stream B — Cyclic durability
A second sample is contaminated with site grit (where relevant), then cycled open-and-closed many thousands of times at a controlled rate before being checked again.
Both streams finish with an ease-of-operation check against measured force limits. A product is certified only when both samples pass every applicable test in their sequence; a single failure in either stream fails the product. Three samples are supplied in total — one per stream plus a reference kept in as-new condition — and all three are labelled and retained until certification is complete.
Tested as it lives
Samples are tested "as received" — no cleaning, lubrication or adjustment beyond what the product's own instructions allow, and if lubrication is permitted it may only be applied before testing begins. Fixtures replicate the real fixing method and shielding, and each sample is set to its most exposed configuration: a chain winder fully extended, a hinge left open, a cylinder tested with its key inserted.
Corrosion resistance — the CR scale
Stream A subjects hardware to neutral salt spray (NSS) in accordance with AS 2331.3.1. Samples are oriented in the chamber the way they sit in service — a sill fitting faces up, a head fitting faces down — so that salt collects where it naturally would. How long a sample must endure that fog determines its grade, from CR1 (not applicable) through to CR10 (extreme), a range spanning 24 hours to a full 1,000 hours of exposure.
| Grade | Descriptor | NSS exposure |
|---|---|---|
| CR1 | Not applicable | no exposure |
| CR2 | Very low | 24 h |
| CR3 | Low | 48 h |
| CR4 | Mild | 72 h |
| CR5 | Moderate | 96 h |
| CR6 | Normal | 168 h |
| CR7 | High | 240 h |
| CR8 | Very high | 480 h |
| CR9 | Severe | 720 h |
| CR10 | Extreme | 1000 h |
CR2 must run continuously. Grades CR3 and above may be paused at 24-hour intervals for inspection — with limited rinsing and a few test operations permitted — provided the sample stays mounted and testing resumes within 30 minutes.
How a pass is judged
After the exposure ends the sample is rinsed, dried and assessed against criteria that differ by finish type:
Decorative finish
No base-metal corrosion visible at 1 m to the naked eye across the rated time; no tarnish, blackening or discolouration; blistering no worse than density 2, size 2.
Protective finish
No visible substrate corrosion at 1 m, allowing up to three small spots per 50 × 50 mm of significant surface — none larger than 2 mm in any direction.
Functional
Once cleaned and dried the product must remain fully operable and deliver its intended function.
What does not count as failure
Salt streaking and deposits that are simply artefacts of the test method are not failures. Neither is tea-staining of stainless steel — a surface discolouration that washes off and doesn't affect integrity — nor internal-component corrosion that leaches onto a visible surface without permanently staining the external finish (for instance reddish rust from steel, or white corrosion from aluminium or zinc alloy).
Cyclic durability — the D scale
Stream B answers a blunter question: how many times can you work this mechanism before it stops behaving? For hardware that lives low in a doorway — within 150 mm of the finished floor, where building dust settles — the test first contaminates the sample. Exposed parts are shaken in a sealed container with 100 g of a dry site-grit mixture, then cycling must begin within four hours.
- 50% Plaster dust (plaster of Paris)
- 25% Cement powder (bagged cement)
- 25% Fine sand (bricklayer's sand)
The rig then holds a sash sized to the maximum dimensions and weight the hardware is rated for, and where the hardware normally closes against a weather seal the rig reproduces that resisting force. Each sash type is driven through a prescribed travel at a controlled rate — the manufacturer picks a speed within the band that avoids overheating the mechanism.
| Sash type | Movement per cycle | Cyclic rate |
|---|---|---|
| Window | 620 mm | 5–30 cycles/min |
| Sliding door | 820 mm | 5–30 cycles/min |
| Swinging sash / louvre | 90% of rated opening | 3–30 cycles/min |
| Folding sash | 90% of opening (4-panel) | 2–30 cycles/min |
The number of completed cycles fixes the durability grade, D1 through D10, from 5,000 to a million cycles. Crucially, each opening type carries its own minimum grade — the floor a product must clear before it can claim to suit that application.
| Grade | Cycles | Minimum required for |
|---|---|---|
| D1 | 5,000 | Windows |
| D2 | 10,000 | Folding doors |
| D3 | 25,000 | Sliding doors |
| D4 | 50,000 | — |
| D5 | 75,000 | Hinged doors |
| D6 | 100,000 | — |
| D7 | 300,000 | — |
| D8 | 500,000 | — |
| D9 | 750,000 | — |
| D10 | 1,000,000 | — |
A grade above the minimum is always acceptable for that opening type. A lockset already tested to the durability sections of AS 4145.2 or AS 4145.3 is deemed to comply.
How a pass is judged
At the end of cycling the sample must not break, must remain fully operable and pass its ease-of-operation check where applicable, and any hardware that relies on friction or resistance to control an opening must still perform that job. The simulated sash weight, the as-installed clearances, the cyclic rate and the cycle count are all recorded in the test report.
Ease of operation — force limits
Durability means little if the hardware becomes stiff to use. Both streams therefore measure the effort needed to operate the hardware, with the rig's driving mechanism disconnected so the fitting can be worked freely. Stream B is checked twice — once before cycling and once after — to confirm operation hasn't degraded. For locksets the torque and force limits follow AS 4145.2; the headline everyday limits are gathered below.
Operating a sash by hand
Force applied at the edge of the sash:
Operating specific hardware
Maximum load to actuate the fitting:
Friction hardware is treated differently
Fittings designed to resist wind purely by friction — casement and awning friction stays — are the exception. Operating a swinging sash held by such a stay may take up to 90 N at the sash edge. And where friction alone controls an open or projected sash, it must hold the sash still against a force of 35 N per m² of sash area, applied perpendicular to the sash face at every opening angle up to 70% of the maximum. All of these figures describe the hardware acting alone, free of any drag from seals or other fittings, and every operating force is recorded for the report.
Certification & labelling
Any hardware claiming or labelled as ICP-002 compliant must be backed by a certificate. The certificate records the code number and date, the test-report reference, the manufacturer's details, a full product description with part numbers, the capacity or size as tested, the product type, and the two performance ratings — corrosion resistance category and cyclic durability grade. The report itself carries a representative drawing so the exact product tested can be uniquely identified.
Where a label is used, it carries a compact version of the same facts. Capacity must be stated for each hardware type and size — for a friction stay, that means both its length and its rated sash weight, not merely the model.
Layout redrawn to KEVOS® house style from the ICP-002 label template — populate each field from the product's own test report.
Fit for purpose considerations
Alongside the graded tests, ICP-002 offers an informative checklist of what "fit for purpose" should mean across a product's nominal life. It's a useful lens for specifiers even where a formal grade isn't the deciding factor.
Structural integrity
Operable without permanent deformation or breakage, and remains serviceable under wind forces up to its rated load.
No secondary damage
Lets the sash operate without damaging any part of the window or door opening around it.
Operation
Moves through its intended functions across the full range of sash loads for which it is recommended.
Security
Provides a level of security suited to its recommended use — referenced where relevant to AS 4145 or another standard.
Noise
Operates without noise beyond what would reasonably be expected of that type of hardware.
Safety
Safe to operate and never a danger to a person during normal use.
The short version
Two ratings tell the story
A CR grade (24 h–1000 h salt spray) and a D grade (5,000–1,000,000 cycles) let any two products be compared on corrosion and endurance the same way.
Separate samples, both must pass
Corrosion and cycling run on different samples; a certificate issues only when both streams — plus their ease-of-operation checks — pass.
Minimums are application-specific
Windows need at least D1, folding doors D2, sliding doors D3, hinged doors D5. A higher grade always qualifies.
It grades hardware, not systems
Weather-tightness and wind resistance of the opening are AS 2047's job. ICP-002 supports comparison, never substitution.
Reading a spec sheet
When you see a fitting marked something like CR6 / D3, read it as: withstands 168 hours of salt spray without substrate corrosion, and survives 25,000 operating cycles while staying smooth to use — enough, on both counts, to suit a sliding door in a typical exposure.