Warping is a common phenomenon found in wood products and is defined simply as a distortion from “flatness”. The loss of bound water in the wood products causes the product to shrink and ‘bow’ or ‘cup’ which occurs if the shrinkage is uneven. Wood warping is an unwanted phenomenon for customers and results in huge losses for New Zealand wood industry in an order of millions of dollars per year. This Masters Project is focused on studying a particular type of wood composite known as Triboard which is manufactured by JNL based in Kaitaia, New Zealand. Triboard is a three-layered panel with a strand board in the core and two MDF fibre layers in the surfaces. JNL has been having issues with warping of this product for some time. In this project, the objectives are to conduct a study to better understand and quantify the factors which cause the Triboard warping, and subsequently to reduce the warping of this product. The experiment was designed to investigate the effects of certain variations in the recipe which were adhesive loading on strands and fibres, moisture content in strands and fibres, hot press temperature and fibre thickness. 27 Triboard samples were made at AICA in New Plymouth, and fibre and strands were supplied JNL, Kaitaia. A base recipe for the control board was used first to make a board of control. Preliminary tries were first conducted to determine the hot press platen temperature in the control recipe which was found to be 1700C. Other parameters in the control recipe were as follows: Board thickness 21 mm (a 3mm MDF layer on each of the outer surfaces, a 15 mm strand layer in the core); MUF solution loading on fibres: 18.2% (solid loading 12%); ––––pMDI solution loading on strands 4.5% (solid loading 3%). Following this, 4 sets of experiments were conducted to investigate effects of resin loading (Set A), moisture content of MDF fibres and strands (Set B), hot press platen temperature (Set C) and thickness of MDF fibre layers (Set D). Moisture content gradient, modulus of elasticity of MDF layers and strand layers, and gradient of strain/stress were measured for each sample board. The criteria for assessing board warping is the difference between maximum and minimum stresses, and the asymmetrical distribution of the stress (quantified as the difference between two surface stresses). Set A had 9 samples each with variations in adhesive loading both in fibres (MUF from 15.2% to 22.8%) and in strands (pMDI from 4% to 5.3%). The results show that the moisture content at the surface layers (5% to 6.5%) were much lower than that in the core layers (9% to 10%). The average moisture content changed from 7.1% to 8.3% depending on the resin loading. Triboard sample A7 with the lowest resin loadings (15.2% MUF, 5.3% pMDI) was observed to have the lowest average moisture content of 7.10% dry basis while sample A3 with highest resin loadings (22.8% MUF, 4.53% pMDI) had the highest moisture content of 8.30%. From the results of large slices, Triboard sample A6 with highest MUF loading on fibres (22.8%) and lowest pMDI loading (4%) was found to be the most stable with the lowest stress difference and most symmetrical stress distribution. For this sample, the difference between the maximum stress and the minimum stress was 5.45 MPa and the asymmetrical stress distribution gave a value of 0.48 MPa. However, sample A9 with 22.8% MUF loading and 5% pMDI loading was the most unstable having a stress difference of 11.66 MPa and stress asymmetry of 6.68 MPa. The same tests were done on small slices in the same way as for large slices to check if they agree. According to strain measurement on small slices, A6 sample maintained the most stable with the least stress difference (9.84 MPa) and most asymmetrical stress distribution (-3.95 MPa). Therefore, it is most likely that the high MUF loading on MDF fibres improves the board stability, but the pMDI loading on the strands within the tested range has less impact. Set B examined the effects of moisture content of fibres (from 10% to 15%) and strands (from 9% to13%) during preparation. Varying moisture in fibres was a very difficult job and so only 6 samples were prepared. The average moisture content of samples for this set was found to increase with added moisture to the fibres and the strands. With increase in fibres and strands moisture contents, moisture contents of the boards at surfaces changed from 4.5% to 6.7% while the core moisture contents changed from 9.5% to 13.8%. The average moisture content changed from 7.18% to 10.19%. Sample B2 with fibre moisture content of 12% and strand moisture content of 9% showed the lowest stress difference (2.83 MPa) and stress asymmetry (2.83 MPa) while sample B6 with fibre moisture content of 15% and strand moisture content of 13% had the most stress difference (29.28 MPa) and the most stress asymmetry (22.88 MPa). This concludes that control of the moisture contents of fibres and strands is very important. Moisture content of strands should be controlled at 9% or lower while maintaining the fibre moisture content at 11-12%. The small sliced samples gave different results however, based on the stress/strain gradient through the board thickness, these results are less reliable due to significant moisture loss during long time and storage and slicing. Set C investigated the effect of hot press temperatures. Results from large slices suggest that average moisture content decreased with increase in hot press temperature. Sample C1 (hot press temperature of 160 0C) had the highest moisture content of 7.47% and sample C3 (pressed at 180 0C) had the lowest average moisture content of 6.71%. Large slices show that the most stable board was sample C3 (with hot press temperature of 180C) which had the maximum stress difference of 5.83 MPa and stress asymmetry of -2.63 MPa. Sample C1 was the most unstable board with maximum stress difference of 13.07 MPa and stress asymmetry of 13.07 MPa. This is probably due to residual stresses being relieved at high temperatures. Once again, the results from small slices do not agree with the above results which are less reliable. Set D examined the effect of varying fibre layer thickness in both top and bottom layers. The thickness of MDF layers was varied between 2-3.5 mm. Results from the large slices show that the average moisture content did not change significantly with varying the fibre thickness from 6.79% to 7.64%. It is found that sample D4 (2 mm MDF fibre layer on the top and 2.5 mm on the bottom) to be the most stable with the maximum stress difference of 5.39 MPa and stress asymmetry of 1.11 MPa, respectively. Sample D6 (2 mm MDF on the top and 3.5 mm on the bottom) was found to be the worst with maximum stress difference and stress asymmetry of both 12.34 MPa. Results from small slices show a slightly different result. Sample D5 (2mm MDF on the top; 3 mm MDF on the bottom) was the most stable with maximum stress difference and stress asymmetry of both 2.27 MPa. Sample D3 (3mm MDF on the top; 3mm MDF on the bottom) was the most unstable board with difference of maximum stress of 21.49 MPa and asymmetric stress of 15.18 MPa. Therefore, it is suggested that the MDF fibre layer on the top should be 0.5 to 1 mm thinner than the MDF layer on the bottom for the 21 mm thick Triboard.