A large-scale fire test was conducted on a compartment constructed from cross laminated timber (CLT). The internal faces of the compartment were lined with non-combustible board, with the exception of one wall and the ceiling where the CLT was exposed directly to the fire inside the compartment. Extinction of the fire occurred without intervention. During the fire test, measurements were made of incident radiant heat flux, gas phase temperature, and in-depth temperature in the CLT. In addition, gas flow velocities and gas phase temperatures at the opening were measured, as well as incident heat fluxes at the facade due to flames and the plume leaving the opening. The fuel load was chosen to be sufficient to attain flashover, to achieve steadystate burning conditions of the exposed CLT, but to minimize the probability of uncertain behaviors induced by the specific characteristics of the CLT. Ventilation conditions were chosen to approximate maximum temperatures within a compartment. Wood cribs were used as fuel and, following decay of the cribs, selfextinction of the exposed CLT rapidly occurred. In parallel with the large-scale test, a small scale study focusing on CLT self-extinction was conducted. This study was used: to establish the range of incident heat fluxes for which self-extinction of the CLT can occur; the duration of exposure after which steady-state burning occurred; and the duration of exposure at which debonding of the CLT could occur. The large-scale test is described, and the results from both the small and large-scale tests are compared. It is found that selfextinction occurred in the large-scale compartment within the range of critical heat fluxes obtained from the small scale tests.