Current research at the University of Canterbury is investigating the performance of a new type of timber floor system made of a timber-concrete composite. This newly proposed timber floor system uses double LVL members connected together with screwed connections to form one larger LVL member. Recent large scale fire tests showed that the joint between these two screwed LVL members opened up during fire exposure. This opening phenomenon causes concerns as the overall charring rate of the joint LVL members is subsequently increased. The main focus of this research, therefore, was to examine the charring rate for different cross sections of single and double LVL members, with different connection types for the double members. The single LVL member examined was 63mm width whereas the double LVL members examined were 90mm and 126mm width. Three connection types were investigated which were nails, screws and glue. Their corresponding charring rates and burning characteristics were examined both in the small furnace provided by the University of Canterbury and in the pilot furnace at the Building Research Association of New Zealand (BRANZ) in Wellington. The overall finding from the small furnace testing shows that the overall average side charring rate for a 30 minute fire exposure was 0.76mm/min; whereas the overall average side charring rate for a 60 minute fire exposure was 0.66mm/min. Moreover for a 30 minute fire exposure, the average bottom charring rates for nail, screw and glue connected double LVL members were 1.00mm/min, 0.83mm/min and 0.83mm/min, respectively. For a 60 minute fire exposure, the average bottom charring rates for screw and glue connected double LVL members were 0.97mm/min and 0.57mm/min, respectively. The nail connected double LVL members experienced the highest bottom charring rate as it suffered the largest bottom separation which allowed the heat to travel into the mid-span resulting in a higher bottom charring rate. Out of these three connection types, the glued connection was the best connection type. Experimental findings were compared with the simulated results generated by the SAFIR finite element program. Experimental findings were also used to modify the spreadsheet design tool which predicts the fire resistance rating of a timber-concrete composite floor under user defined load conditions and floor geometries.