pvlib.soiling.hsu takes a tilt parameter representing the same thing we normally call surface_tilt:

I don't see any good reason for this naming inconsistency (I suspect tilt just got copied from the matlab implementation) and suggest we rename the parameter to surface_tilt with a deprecation.

Also, the docstring parameter type description says it must be float, but the model's reference explicitly says time series tilt is allowed:

The angle is variable for tracking systems and is taken as the average angle over the time step.