Sitting on a bench with the balls of the feet on a low step and a 20 kg dumbbell resting across each knee, held there by the hands. The ankle moves from 12° of stretch below the step to 26° of raise.
Load in the model: 20 kg each hand. Balance and joint demand re-solve as you change it.
Faults grow with knee bend and apply to lifts standing on the floor. The fault check flags a knee caving more than 10° inside the line of the toes and a load split beyond 60/40. Knee cave shows best in the front camera.
Femur changes keep standing height fixed by shortening or lengthening the torso. Longer femurs force more forward lean in any squat.
Your quota per set: 10 reps at 20 kg each hand, stopping 2 reps short of failure.
One hard set earns 70 legs points.
Soleus (the bent knee takes gastrocnemius out), gastrocnemius.
One 3D skeleton drives both views. Segment lengths and widths are Winter's anthropometric ratios of height; segment masses and centre-of-mass positions follow Dempster's data. For free-standing lifts the whole-body angle is found numerically so the combined centre of mass of lifter and load sits over the middle of the foot. Arms and legs use two-link inverse kinematics with a pole direction. Rep timing follows a minimum-jerk profile.
Balance is quasi-static, so ballistic lifts are shown slowed down. Cable pull and pad support forces are not modelled. The spine is one rigid segment and joints have no range-of-motion limits or muscle forces.