As the blades become non- perpendicular to the air flow, the swept area is reduced and it follows the cosine of the angle difference between perpendicular and the tilt.
This lowering in apparent swept area relative to the flow of the air reduces the power.
Actually it's a tad different:
As you become non-perpendicular the component of the airflow along the axis is lower. You've slowed the wind seen by the turbine by that cosine factor.
The available power is the CUBE of the airspeed, so the reduction in potential power is 1 - cos(angle)^3.
However it's not that simple for two reasons:
- A windcharger does not load the turbine optimally, so the output power is not a cube function of the wind. All that matters is current, which is proportional to torque (and to (gen voltage - (battery voltage + diode drop))/series resistance). So it's pretty close to a first order function of (wind - cutin wind speed).
- The cross-axis component of wind interacts with the blades in a complex way.
Still, a 5 degree tilt probably doesn't make enough difference to measure with ordinary test equipment.