The samples are assembled on the testbed in accordance with how they are intended to be used, with the ends and sides sealed to minimise loss of air pressure. Measuring devices for deflection are placed at static points (fasteners) and pan centre at mid-span or as required to measure maximum deflection.
The air pressure is raised incrementally, and the point of deflection failure noted. With pierce-fastened products, signs of imminent deformation are observed. When product nears deformation failure, pressure may be periodically backed off to allow loosening of a fastener to inspect the integrity of cladding profile underneath. The pressure at these points should be noted. The maximum serviceability load test result is the recorded pressure level at the last observation before the onset of failure or the point before which the pressure level recorder showed a sudden small drop signifying profile movement.
More than one serviceability result may be recorded from separate fasteners during a single test if, at the discretion of the testing engineer, the loads imposed on subsequent failure points are equivalent to those imposed on the initial failure point
If lap screws are added for the purpose of attaining failure pressures by decreasing air leakage, these must not be positioned closer than 0.5 m to the nearest purlin.
Clip fastened sections that de-clip at the central rib or de-index on a lap but still resist increased pressure, represents a serviceability failure.
Load is then increased to establish the strength load result. This may be when pierce-fastened cladding pulls fasteners through the profile or clip fastened profile de-clips. Both these actions result in sudden and permanent loss of air pressure and terminate the test. The load at which this occurs, or maximum load achievable due to air leakage, is recorded.
If strength load limits cannot be attained by the testing apparatus, a load of at least 2.0 times the serviceability load must be imposed to demonstrate that the sample failure load is governed by serviceability rather than by strength. If serviceability/strength ratios exceed 2.0 for at least two different spans, it may be assumed that the ratio will be similar on all spans.
Fastener pull-out from testbed purlins is not recorded as a failure as this is independent of the profile strength and is relevant only to the condition of the purlin material on the testbed. Such fasteners may be replaced or repositioned and the test recommenced.