Testing — salamander-v2

Test surfaces

Three surfaces planned, in order of difficulty:

1. Linoleum (lab floor) — low friction, flat. The baseline.
2. Carpet (living room) — high friction, compliant. Legs stall.
3. Concrete (garage) — high friction, rigid. Not attempted yet.

Linoleum results

Forward gait at 0.3 body-lengths per second. Stable for ~2 minutes before the spine drift accumulates and the body curves left. The turning radius with spine actuation is approximately 30cm, which is acceptable.

Battery life under continuous walking: 14 minutes on a 500mAh 2S pack. The servos draw more than expected at stall — the front pair hits stall on every recovery stroke.

Carpet results

Failed. The front legs cannot overcome the carpet friction on the recovery stroke. The gait controller detects the phase drift but cannot correct because the servos are at torque limit. Need either stronger servos or a different gait strategy for high-friction surfaces.

Next tests

• Linoleum with the new vertebrae spine (once printed)
• Measure power draw per servo at various loads
• Try a tripod gait on carpet (fewer legs moving at once = more torque per leg)