Timber is increasingly used in buildings because of its favourable embodied carbon credentials compared to traditional materials such as steel and concrete. However, due to it combustible nature, encapsulation products are frequently used to prevent or limit the timbers contribution to a fire. These encapsulation products introduce additional weight, cost and carbon to project, which limit the benefit of timber construction. The performance of these products in natural fires and the relative fire severity experienced by products placed near the ceiling and the floor has received little attention. Here we present data on the performance of encapsulation applied to the timber ceiling and floor elements within the CodeRed experiments – a series of large-scale timber compartment experiments with varying ventilation and extent of exposed timber. Compartment temperatures were measured at the ceiling and near the floor, indicating a lower temperatures at floor level; the time temperature curve ‘seen’ by the floor encapsulation is dependent on whether or not there is residual glowing embers on the floor. Both the 25 mm calcium silicate encapsulation to the floor, when located below the wood crib and when adjacent to it, and the three layers of 12.5 mm gypsum fibreboard board applied to the ceiling were shown to be adequate in preventing the ignition of the underlying timber. However, smouldering was observed to sustain remote to encapsulation, but eventually spread beneath the encapsulation which facilitated continued smouldering. This highlights that smoulder can progress behind encapsulation. A 1D Finite Difference Model (FDM) was used to explore the temperature development of timber surfaces beneath simulated encapsulation details using the gas temperatures measured near the ceiling and floor. Converting the modelled peak temperatures to time equivalent times revealed an 23% average reduction in time equivalent fire severity near the floor level for natural fires with a severity equivalent to a 30 min standard fire. When considering greater fire severities, this reduction remained similar, indicating that applying the same fire protecting rating to the floor as to the ceiling is likely overly conservative. The study represents the first step in quantifying a more refined protection required at the floor and to the ceiling; with further research it may be possible to justify a reduction in the required fire protection performance of floor level encapsulationfor long duration fires compared to that on the ceiling, taking into account the observed distribution of compartment gas temperatures in large compartments. Ultimately this offers an opportunity to reduce the embodied carbon, costs, and weight of the structural design.