Theory and Method
Direct shear and pullout look similar from outside the machine, but they measure different mechanisms and neither result substitutes for the other. Direct shear forces sliding along a defined plane. The geosynthetic is fixed to a rigid substrate or sandwiched between soil layers, normal stress is applied through the top platen, and the box halves displace horizontally while the load cell records resistance. Running the test at three or more normal stresses produces a failure envelope whose slope gives the interface friction angle and whose intercept gives adhesion. Those two numbers drop straight into a slope stability calculation.
Pullout mobilizes something else. The reinforcement is buried in soil, gripped at one end, and drawn out horizontally while the surrounding soil stays put. Friction develops on both faces of the material, and for geogrids the transverse ribs also push against the soil ahead of them, generating passive bearing resistance that a sheet product cannot develop. That bearing component often dominates, which is exactly why a geogrid anchors far better than its surface area alone suggests, and why pullout resistance cannot be derived from an interface friction angle. The output is a resistance versus displacement curve that tells a designer how much embedment length is needed to reach the required anchorage force.
Box size governs how much the result can be trusted. At 300 × 300 mm the specimen spans several apertures of a typical geogrid, so the measurement reflects the product's real geometry rather than a single junction, and the sidewalls sit far enough from the shear plane to limit boundary friction. Servo normal loading matters for the same reason: dense granular fill expands as it shears, and a system that cannot follow that movement lets normal stress climb and overstates strength. The DSP holds the applied pressure steady through dilation, and the 0.02 to 3 mm/min rate range covers everything from slow drained shear in silty fill to faster runs on clean sand.