When To Use Ray Aiming

This article will talk about what Ray Aiming is, when to use it and some recommendations if you get errors.

Summary

The tl;dr summary:

• If Pupil Aberrations are less than 2%, do not use Ray Aiming.?
• If Stop is in Object Space, do not use Ray Aiming.
• If Ray Aiming is needed, always use Float By Stop Size and Paraxial Ray Aiming.

Defining a Ray

A ray is fully defined by knowing the starting location, the initial direction and the wavelength.?If the first two values are fully defined using XYZ positions and LMN direction cosines, then we don’t have to do any conversion for the ray trace algorithm.?However, all systems are defined using Hx/Hy normalized field coordinates and Px/Py normalized pupil coordinates.?From these 2 sets of points in space, OpticStudio needs to calculate the XYZ starting location and the LMN initial direction cosine.

For all systems in the OpticStudio GUI, the XYZ starting position is well defined with the Hx/Hy starting pair; if it’s a finite conjugate, the XYZ are explicitly defined in the Lens Data Editor and the Field Data Editor and if it’s an infinite conjugate system the XYZ are on a plane tangent to the chief ray in object space.?

The tricky situation comes with figuring out the initial direction cosines. With the Hx/Hy defined, OpticStudio needs a second plane located somewhere else in the optical system to calculate another set of "target points" on and then determine what LMN direction cosine connects the two sets of points.?

With no Ray Aiming, OpticStudio uses the Paraxial Entrance Pupil as this second plane and as long as there are very “low mapping differences” (pupil aberrations) between the Paraxial Entrance Pupil and the actual Stop, using the Entrance Pupil is fine.?These pupil aberrations are defined as a percent difference of a real ray hitting the same normalized location in both the Entrance Pupil and the Stop (i.e., the chief ray goes through the center or the marginal ray hits the edge).?You can see these aberrations by clicking Analyze > Aberrations > Pupil Aberrations:

If the total pupil aberrations are less than about 1%-2%, then ray aiming is not needed.

Ray Aiming

If ray aiming is needed, then OpticStudio uses an 2 step iterative algorithm to change the initial LMN direction cosine until the ray hits the correctly defined Px/Py position in the actual Stop (and not the Entrance Pupil); the initial XYZ is also modified for infinite conjugate systems as long as the ray still falls on the plane tangent to the chief ray.?The first step is to find a single "best guess" ray that mathematically traces from the Object Plane to the Stop. If this "best guess" ray makes it to the Stop, then OpticStudio will slightly iterate this ray until it makes it to the correct Px/Py location. If this "best guess" ray does not make it to the Stop, then a more robust search algorithm is used to try to determine the direction cosine for any ray that can make it to the Stop. If Ray Aiming fails, it's typically because OpticStudio cannot efficiently find a ray that makes it from Object Plane to the Stop.

The difference between Paraxial Ray Aiming and Real Ray Aiming is how the physical stop size is set if the Aperture Type is set to anything other than Float By Stop Size.?Since most first-order properties use the Paraxial Entrance Pupil (F/#, NA, efl, m,?etc), Paraxial Ray Aiming uses paraxial rays to determine the size of the Stop; this option makes sure that first-order properties are still calculated using the Paraxial Entrance Pupil.?Real Ray Aiming will use real rays rather than paraxial rays to find the size of the Stop.?For both algorithms, once the size of the Stop is determined, the same algorithm is used to determine the initial XYZ and LMN value and real rays are traced through the system for aberration analysis.?However, Real Ray Aiming is about 2x-8x slower and fails in more instances than Paraxial Ray Aiming.

Ray Aiming Rules of Thumb

A few things to consider when using Ray Aiming:

• Ray Aiming is not needed if the Pupil Aberrations are less than 1%-2%
• You should always set the Aperture Type to Float By Stop Size if Ray Aiming is turned on (the algorithm needs to have something to aim at)
• If you need Ray Aiming, start with Paraxial Ray Aiming, check the Pupil Aberrations and if the aberrations are now less than 1%-2%, you don’t need Real Ray Aiming

How to Fix Errors

If there are errors (for advanced users and compiled by a computational physicists at Zemax):

• If using Real Ray Aiming, switch to Paraxial Ray Aiming
• If pupil shifts are off, try turning them on
• If pupil shifts are on, turn them off and set the values to:
• 0 if the values are small and make sense
• To a reasonable value if one of them doesn’t make sense (i.e., the Z Shift is past the Image Plane)
• Change the Semi-Diameter of the first surface with optical power (there is a built-in grid search for determining the correct XYZ and LMN values to use)
• Make sure it isn’t some absurd value
• Set it to Fixed in the Clear Semi-Diameter solve box
• Try to change the size of the Semi-Diameter by either a factor of 2x or 0.5x
• Set the aperture type to Float by Stop Size
• Use a Field Type defined in object space
• Toggle the Ray Aiming Cache off/on
• Turn on Robust Ray Aiming?