Fifteen stems of black spruce (Picea mariana (Mill.) BSP) coming from the AbitibiTemiscamingue region in Canada were crosscut into three sections: bottom, middle, and top logs. Logs were fragmented producing three faces with a chipper-canter using three cutting widths (CWs) of 12.7, 19.1, and 25.4 mm. Chip dimensions were assessed by thickness, width, and length (Domtar and Williams classifications). Knot characteristics (total knot number [TKN] and area [TKA]) were assessed in the three cant faces. Growth ring attributes (earlywood density, latewood density [LWD], ring density, earlywood proportion, ring width, and rings per mm [R/mm]), mechanical properties (shear, splitting, modulus of elasticity (MOE), and modulus of rupture (MOR) in bending), and basic density were evaluated on samples obtained within each CW area. The results showed that most of these wood attributes were affected by the log position in the stem and/or CW. Weighted mean chip thickness (WCT) and chip size distributions were significantly affected by log position and CW. WCT increased as CW increased. WCT variation with height could be principally associated with number and size of knots within the stem. However, greater taper in the bottom logs produced thicker chips. Multiple linear regressions showed that CW, TKN, LWD, and TKA were significant predictors of WCT. Also, chip thickness distribution was affected primarily by TKA, cutting height, and LWD, whereas the width and length distribution was mainly affected by R/mm, TKN, and MOE. Chip size variation was to some point determined by knot characteristics, bending properties, growth ring width, and wood density of the raw material. Results showed the potential benefits of classifying logs in wood yards and better controlling the raw material attributes in sawmills. If CW is combined with knowledge of the raw material, chip dimensions can be adjusted using other fragmentation parameters to increase chip size uniformity.
