0 to +90 degree limits on axis tilt is SAT

[markcampanelli]

I am wondering if the 0 to +90 degree limit on axis tilt is correct for SAT calculations in pvlib.tracking.singleaxis. I think this differs from the -90 to +90 degree limit on the analogous cross-axis tilt. It seems that some geometries cannot be represented if axis tilt is restricted to 0 to +90 degrees. Likewise, I think pvlib.tracking.calc_axis_tilt can produce negative axis tilt values without raising an error, much like pvlib.tracking.calc_cross_axis_tilt.

See:

• pvlib.tracking.calc_axis_tilt
• pvlib.tracking.calc_cross_axis_tilt
• pvlib.tracking.singleaxis

[williamhobbs]

This might be related, and have some relevant discussion: #1976

[kandersolar]

It seems that some geometries cannot be represented if axis tilt is restricted to 0 to +90 degrees.

Can't these geometries be represented by keeping a positive axis_tilt but rotating the axis_azimuth by 180? (notwithstanding the issues about consistency across an array; I am just interested in the geometry here)