Skew Ray Tracing and Error Analysis of Optical Lens With Cylindrical Boundary Surface - Liao Te Ten, Chen Shih Hung, Chen Kuo Ying, Chen Chun Ta

SKEW RAY TRACING AND ERROR ANALYSIS OF OPTICAL LENS WITH
CYLINDRICAL BOUNDARY SURFACE

Te-Tan Liao1, Shih-Hung Chen2, Kuo-Ying Chen3, Chun-Ta Chen4
1
Department of Mechanical Engineering, Far East University, Tainan, Taiwan, R. O. C.
E-mail: [email protected]
2
Department of Automation and Control Engineering, Far East University, Tainan, Taiwan, R. O. C.
3
Department of Mechanical Engineering, Far East University, Tainan, Taiwan, R. O. C.
4
Department of Mechanical and Automation Engineering, Da Yeh University, Changhua, Taiwan, R. O. C.
Received February 2008, Accepted January 2009

No. 08-CSME-07, E.I.C. Accession 3045
ABSTRACT
This study applies computational geometric algebra based on a 4 | 4 homogeneous
transformation matrix and Snell’s law of geometrical optics to analyze skew rays and the errors
of a light ray’s path as it passes through a cylindrical lens. The author addresses two important
topics: (1) the determination of the direction of a reflected or refracted ray by Snell’s law and (2)
the expression of the combination of two principal sources of light path error using error
analysis. In topic (2), one of the sources is the translational errors Ddix , Ddiy , and Ddiz and the
rotational errors Dvix , Dviy , and Dviz that determine the deviation of the light path at each
boundary surface, while the other source is the differential changes induced in the incident point
position and the unit directional vector of the refracted/reflected ray as a result of differential
changes in the position and unit directional vector of the light source. The methodology
presented in this study provides a comprehensive and robust approach for evaluating the error
of a light ray path as it passes through a cylindrical lens.
LANCER DE RAYONS OBLIQUES ET ANALYSE DE LENTILLES OPTIQUES À

SURFACE CYLINDRIQUE FRONTIÈRE
RÉSUMÉ
Cette étude met en application l’algèbre géométrique de calcul basée sur une matrice de
transformation homogène 4 X 4 et la loi de Snell sur l’optique géométrique pour analyser les
rayons obliques, et les erreurs de trajectoire d’un lancer de rayons lumineux au moment de
passer à travers une lentille cylindrique. L’auteur s’intéresse à deux points importants : (1) la
détermination de la direction d’un rayon réfléchi ou réfracté selon la loi de Snell, et (2)
l’expression de la combinaison de deux sources principales d’erreur de trajectoire de la lumière
en utilisant la méthode d’analyse d’erreur. Au sujet du point (2), une des sources d’erreur ce sont
les erreurs de translation Ddix , Ddiy , et Ddiz les erreurs de rotation Dvix , Dviy , et Dviz qui
déterminent la déviation de la trajectoire de la lumière à chaque surface frontière. Tandis que
l’autre source d’erreurs ce sont les changements différentiels provoqués dans la position du
point d’incidence et le vecteur directionnel du rayon réfléchi ou réfracté, résultant des
changements différentiels dans la position et le vecteur différentiel de la source lumineuse. La
méthodologie présentée dans cette étude, propose une approche compréhensive solide pour
l’évaluation de l’erreur de trajectoire d’un rayon lumineux au moment de passer à travers une
lentille cylindrique.
Transactions of the Canadian Society for Mechanical Engineering, Vol. 33, No. 2, 2009 297
1. INTRODUCTION
Currently, the ray tracing method is the core method used in many optical design tools and
analysis software. The evaluation of the performance of an optical system during its theoretical
design stage requires the ability to determine the paths of the light rays as they undergo
reflection and refraction at the boundaries of the various optical elements within the system.
The light path can be determined using a ray tracing technique in which the optical laws of
reflection and refraction are systematically applied at each boundary encountered by the light
ray [1,2]. In general, the light rays in an optical system can be classified as either axial rays,
meridional rays, or skew rays. Axial rays and meridional rays can be traced using relatively
simple trigonometric formulas or even graphically if a low precision is acceptable. Meridional
rays in the paraxial region of an optical system can also be approximately traced via the
successive application of matrix production [3]. A skew ray, which is the most general ray, is
significantly difficult to trace. Nevertheless, skew ray tracing is essential while modeling an
optical system and evaluating its performance. Consequently, when designing optical systems, it
is conventional to trace the rays using actual light sources and the optical components being
used. To facilitate the tracing of skew rays, Lin [4] reformulated the traditional optical laws of
reflection and refraction in terms of revolution geometry. Subsequently, he conducted a
sensitivity analysis based on a method of skew ray tracing to determine the changes in a light
ray path as it crossed the boundary between the different media.
Ray tracing provides a powerful technique for analyzing the performance of optical lenses
and is an essential task in the analysis and design of optical systems. The traced rays generally
include those that start at a given set of object points and pass through a given set of points on
the aperture stop [5]. In the differential ray tracing process, the effects of each optical
component are evaluated by differentiating the equations related to the ray configuration of the
rays before and after their transformation at a component surface [6,7]. Such ray tracing
approaches enable the sensitivity of an optical system to design or manufacturing flaws to be
assessed by correlating the differential changes of the reflected or refracted rays with the
differential changes in the incident rays. Thus, the effect of each optical boundary surface on the
overall ray path within the optical system can be systematically examined. Current commercial
software for the analysis and design of optical systems use finite difference (FD) approximation
methodology to estimate the gradient matrix of a ray with respect to the system variables.
However, FD estimates are intrinsically inaccurate and are subject to a gross error when the
denominator is excessively small with respect to the numerator. Lin and Tsai [8] avoid these
problems and determine these gradients by the application of Snell’s law. They provide
information on the background and basics for determining the first-order gradients of skew
rays of optical systems. Using these gradients, the differential vector of any ray can be estimated
by determining the product of the developed gradient matrix and differential changes in the
system variables. Furthermore, sensitivity analysis enables the construction of fundamental
aberration functions, which greatly simplify the optical design task [9,10]. Lin and Liao [11]
demonstrate the effectiveness of the sensitivity analysis approach by using a Jacobian matrix
and the Newton-Raphson method to determine the path of the chief ray in a binocular
stereovision system. Sensitivity analysis also enables the orientation of an image to be
accurately determined. For example, by considering a prism design for illustration purposes,
Tsai and Lin [12] applied the results of sensitivity analysis to construct a merit function
describing the change in the orientation of an image as it was successively reflected/refracted at
a series of flat boundary surfaces.
In many ways, the procedure used to design an optical system resembles that employed to
design a mechanical machine tool. In the same way that optical systems comprise a structured
series of optical elements, NC machine tools comprise a series of interconnected links and
joints, each with its own unique mechanical properties and resolution characteristics. When
designing a machine tool system, it is essential to identify the potential sources of error within
the system and to clarify their individual and combined effects on the overall system
performance so that the quality of the machined products can be reliably determined. In an
earlier study [13], Tlusty showed that the overall error of an NC machine tool varies as a
function of the combined effects of the individual errors in each of its six degrees of freedom.
Ferreira and Liu [14] developed a model based on three rotational errors (DCi , DYi , and DWi )
and three translational errors (DXi , DYi , and DZi ) to enable the performance of a machine tool
to be systematically examined. Considering the case of a jig-boring machine, Schultshik [15]
applied these error definitions to investigate the combined effect of the various components of
the volumetric accuracy of a machine tool on its machining precision.
The precision of the components produced by a machine tool system is dependent on the
quality and configuration of each of its links and joints. Similarly, the image quality of an
optical system is determined by the errors that inevitably exist in the fabrication and assembly
of its individual optical components [16]. Therefore, by employing an approach similar to that
used by Ferreira and Liu in analyzing the geometric errors of a machine tool [14], the current
study develops a skew ray sensitivity analysis model based on six sources of light path errors,
i.e., three rotational errors and three translational errors, to analyze the errors induced in a ray’s
light path as it is reflected and/or refracted by a succession of cylindrical boundary surfaces
within an optical system. Significantly, in determining the deviation of the light path at each
boundary surface, the proposed model incorporates an error matrix that takes account not only
the effects of the rotational errors and translational errors but also the differential changes
induced in the incident point position and unit directional vector of the refracted/reflected ray
as a result of differential changes in the position and the unit directional vector of the light
source. The present study achieves a more efficient cylindrical boundary skew ray tracing by the
use of Snell’s laws formulated as homogeneous transformation matrices, as described in Section
2. Error analysis is presented in Section 3. Numerical simulation of the errors of a refractive
ray’s light path as it passes through an optical element with a cylindrical boundary surface is
presented in Section 4. The conclusion is presented in Section 5.
In the derivations performed in this paper, the position vector Pix izPiy jzPiz k is written in
T
the form of the column matrix g Pi ~½ Pix Piy Piz 1 , where the pre-superscript ‘‘g’’ of the
g
leading symbol Pi indicates that the vector is referred with respect to the coordinate frame
(xyz)g . Furthermore, given a point g Pi , its transformation k Pi is represented by the matrix
product k Pi ~k Ag g Pi , where k Ag is a 4|4 matrix defining the position and orientation
(referred to hereafter as the configuration) of the frame (xyz)g with respect to another frame
(xyz)k [17]. The same notation rules are also applied to the unit directional vector
g T
Vi ~½ Vix Viy Viz 0 . Note that for the vectors referred with respect to the world frame
(xyz)0 , the pre-superscript ‘‘0’’ is omitted for convenience.
2. SKEW RAY TRACING FOR CYLINDRICAL LENS

A fundamental feature of typical optical components is that they posses boundary surfaces that
can be described by revolution geometry. Therefore, the first step in the determination of a typical
optical system’s skew ray path is the establishment of suitable revolution boundary surfaces.
A boundary surface i ri at which reflection and refraction processes occur, can be obtained by
the rotation of its generating curve ½ x(bi ) 0 z(bi ) 1 T in the xi zi plane about its optical zi
axis. i.e.,
2 32 3 2 3
Cai {Sai 0 0 x(bi ) x(bi )Cai
6 Sai Cai 0 0 7 6 7 6 7
i
ri ~ 6 76 0 7~6 x(bi )Sai 7 (1)
4 0 0 1 0 54 z(bi ) 5 4 z(bi ) 5
0 0 0 1 1 1
where C and S denote the cosine and sine functions, respectively. As shown in Fig. 1, i ri for an
optical element with cylindrical boundary surface can be obtained by rotating the generating
curve i qi ~½ Ri 0 hi 1 T in the xizi plane about the zi axis, i.e.,
2 32 3
Cai {Sai 0 0 Ri
6 Sa Ca 0 0 7 6 7
i
ri ~Rot(zi ,ai )i qi ~6 i i 76 0 7~½ Ri Cai Ri Sai hi 1 T , (2)
4 0 0 1 0 5 4 hi 5
0 0 0 1 1
where Rot(zi ,ai ) is the rotation transformation matrix about the zi axis. Equation (2) provides a
generic expression for parametrizing the boundaries of optical elements with cylindrical surfaces
in terms of Ri and the polar angular position ai . The unit normal vector i ni along i ri is given as
i
ni ~si ½ Cai Sai 0 0 T , (3)
where si is set to +1 or –1 so that the cosine of the incident angle has a positive value, i.e.,
Chi w0.
Equations (2) and (3) give the parametric expression of i ri and i ni relative to frame (xyz)i built
with respect to the i-th cylindrical boundary surface. However, many derivations in this paper
are built relative to the world frame (xyz)0 . Therefore, we need the relative configuration of
world frame (xyz)0 with respect to frame (xyz)i , which is given as the following 4|4
homogeneous transformation matrix:
Fig. 1. Surface geometry and skew ray tracing for cylindrical optical element.
2 3
Iix Jix Kix tix
6 I Jiy Kiy tiy 7
i
A0 ~Ai0 ~6 iy
4 Iiz
7, (4)
Jiz Kiz tiz 5
0 0 0 1
where i A0 describes the position and orientation of the world frame (xyz)0 with respect to the
T T T
frame(xyz)i . Vectors ½ Iix Iiy Iiz 0 , ½ Jix Jiy Jiz 0 , and ½ Kix Kiy Kiz 0 describe
the orientations of the three unit vectors of frame (xyz)0 with respect to the frame (xyz)i . The
T
vector ½ tix tiy tiz 1 is the position vector of the origin of frame (xyz)0 with respect to
frame (xyz)i (see Fig. 2) [17].
The unit normal vector of the cylindrical boundary surface referred with respect to frame
(xyz)0 , i.e. ni , can be obtained by transforming i ni to frame (xyz)0 , i.e.,
2 3 2 3
nix Iix Cai zIiy Sai
6 niy 7 0 i 6 7
ni ~6 7~ Ai ni ~(i A0 ){1i ni ~si 6 Jix Cai zJiy Sai 7 (5)
4 niz 5 4 Kix Cai zKiy Sai 5
0 0
T
In Fig. 1, a light ray originating at point Pi{1 ~½ Pi{1x Pi{1y Pi{1z 1 and directed
T
along the unit directional vector Vi{1 ~½ Vi{1x Vi{1y Vi{1z 0 is reflected/refracted at the
cylindrical medium boundary surface i ri . The parametric form of any intermediate point
(denoted as P0i ) along the incident ray is
P0i ~½ Pi{1x zVi{1x l Pi{1y zVi{1y l Pi{1z zVi{1z l 1 T , (6)
where l§0 is the magnitude of vector Pi-1 P0i . The parameter l~li at which P0i hits i ri at the
incident point
Pi ~½ Pi{1x zVi{1x li Pi{1y zVi{1y li Pi{1z zVi{1z li 1 T (7)
is given by equating i ri to i P0i (i P0i ~i A0 P0i is the transformation of P0i to frame (xyz)i ):
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
{Di + Di 2 {Hi :Ei
li ~ : (8)
Hi
Fig. 2. Physical meaning of configuration matrix i A0 .
Here, Hi , Di and Ei are given by
Hi ~G2ix zG2iy , (9)
Di ~Gix (Fix ztix )zGiy (Fiy ztiy ), (10)
Ei ~(Fix ztix )2 z(Fiy ztiy )2 {R2i : (11)
Where Fix ~Iix Pi{1x zJix Pi{1y zKix Pi{1z , Fiy ~Iiy Pi{1x zJiy Pi{1y zKiy Pi{1z , Fiz ~Iiz Pi{1x z
Jiz Pi{1y zKiz Pi{1z , Gix ~Iix Vi{1x zJix Vi{1y zKix Vi{1z , Giy ~Iiy Vi{1x zJiy Vi{1y zKiy Vi{1z ,
Giz ~Iiz Vi{1x zJiz Vi{1y zKiz Vi{1z .
The ambiguous sign of the root term in Eq. (8) indicates two possible points of intersection of
the ray with a cylindrical boundary surface. Only one of these points is required in the current
ray tracing analysis, i.e., the initial incident point, and the appropriate sign must therefore be
chosen. The parameters ai and hi of the incident point on the cylindrical boundary surface are
crucial in determining the subsequent path of the refracted ray and are defined as follows:
ai ~arctan(ri ,si ), (12)
hi ~Fiz zli Giz ztiz , (13)
where si ~Fix zli Gix ztix , ri ~Fiy zli Giy ztiy .

To trace the reflected/refracted ray at the cylindrical boundary surface, we still need the
incidence angle hi given by
Chi ~{VTi{1 :ni ~si (Gix Cai zGiy Sai ): (14)
The refraction angle hi between these two optical media must satisfy Snell’s law:
ji{1
Shi ~ Shi ~Ni Shi , (15)
ji
where ji{1 and ji are the indices of media i – 1 and i, respectively. Ni ~ji{1 =ji is the relative
refractive index of medium i – 1 with respect to medium i. The common unit normal vector mi
of ni and Vi{1 is given by
mi ~½ mix miy miz 0 T ~ni |Vi{1=jni |Vi{1 j~ni |Vi{1=Shi : (16)
According to the reflection (refraction) law of optics, the refracted (reflected) unit directional
vector Vi (Vi ) can be obtained by rotating ni about mi at an angle hp ~p{hi (hp ~hi ). This leads to
T
Vi (Vi )~½ Vix Viy Viz 0
2 32 3
m2ix (1{ChR )zChR miy mix (1{ChR ){miz ShR mix miz (1{ChR )zmiy Shp 0 nix
6 7
6 miy mix (1{ChR )zmiz ShR m2iy (1{ChR )zChR miz miy (1{ChR ){mix ShR 0 76 niy 7 (17)
~6 766
7
7:
6 7
4 mix miz (1{ChR ){miy ShR miz miy (1{ChR )zmix Shp m2iz (1{ChR )zChR 0 54 niz 5
0 0 0 0 0
Further simplification of Eq. (17) is possible by utilizing the equation Shi mi |ni ~(ni |Vi{1 )|
ni ~Vi{1 {(nTi :Vi{1 )ni ~Vi{1 zni Chi (see Eq. (16)) and Snell’s law Shi ~Ni Shi , resulting in:
2 3 2 3 2 3
Vix nix ChR z(niz miy {niy miz )ShR nix ChR zNi (Vi{1x znix Chi )
6 Viy 7 6 niy ChR z(nix miz {niz mix )ShR 7 6 niy ChR zNi (Vi{1y zniy Chi ) 7
Vi (Vi )~6 7 6 7 6 7
4 Viz 5~4 niz ChR z(niy mix {nix miy )ShR 5~4 niz ChR zNi (Vi{1z zniz Chi ) 5: (18)
0 0 0
From Eq. (18), the refracted unit direction Vi (hp ~p{hi ) is expressed as a unit vector as
2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3
2 3 {nix 1{N2i z(Ni Chi )2 zNi (Vi{1x znix Chi )
Vix 6 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 7
6 Viy 7 6 6 {n 1{N 2
z(N Ch )2
zN (V zn Ch )
7
7
Vi ~46 7 ~6 iy i i i i i{1y iy i 7, (19)
Viz 5 6 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 7
4 {niz 1{N2 z(Ni Chi )2 zNi (Vi{1z zniz Chi ) 5
0 i
0
while the reflected unit direction Vi (hp ~hi and Ni ~1) is given by
2 3 2 3
Vix Vi{1x z2nix Chi
6 V 7 6 Vi{1y z2niy Chi 7
Vi ~6 iy 7 6
4 V 5~4 Vi{1z z2niz Chi 5:
7 (20)
iz
0 0
Following refraction (reflection), the light ray proceeds with point Pi as its new point of
origin and Vi as its new unit directional vector. We have illustrated how the path of the reflected
or refracted ray relative to a single cylindrical boundary surface can be found. We can apply this
approach successively to trace rays relative to a series of optical elements that possess only
cylindrical boundary surfaces.
3. ERROR ANALYSIS FOR CYLINDRICAL LENS

In optical systems, errors inevitably occur between the designed position and orientation of
the optical elements and their actual position and orientation. In analyzing these errors, the
relative positions and orientations of (xyz)0 with respect to the ideal frame (xyz)i and the actual
frame (xyz)a can be expressed, respectively, as
2 3
Iix Jix Kix tix
6 Iiy Jiy Kiy tiy 7
i
A0 ~ 6 7
4 Iiz Jiz Kiz tiz 5, (21)
0 0 0 1
2 3
Iax Jax Kax tax
6 Iay Jay Kay tay 7
a
A0 ~6
4 Iaz
7: (22)
Jaz Kaz taz 5
0 0 0 1
As shown in Fig. 3, the position and orientation errors of any element with a cylindrical
boundary surface within the optical system can be described in terms of three translational
errors of the origin of frame (xyz)a , i.e., Ddix , Ddiy , and Ddiz , and three rotational errors of the
Fig. 3. Definition of translational and rotational errors of coordinated frame of the cylindrical
boundary surface.
three axes of frame (xyz)a with respect to frame (xyz)i , i.e., Dvix , Dviy , and Dviz . The overall
effect of these six errors can be mathematically expressed using a matrix i Aa of the form
i
Aa ~Trans(Ddix ,Ddiy ,Ddiz )Rot(z,Dviz )Rot(y,Dviy )Rot(x,Dviz )
2 3
CDviz CDviy CDviz SDviy SDvix {SDviz CDvix CDviz SDviy CDvix zSDviz SDvix Ddix
6 SDv CDv SDviz SDviy SDvix zCDviz CDvix SDviz SDviy CDvix {CDviz SDvix Ddiy 7 (23)
6 iz iy 7
~6 7:
4 {SDviy CDviy SDvix CDviy CDvix Ddiz 5
0 0 0 1
Since in an optical system, the translational and rotational errors are small, Eq. (23) can be
approximated by the first-order Taylor series expansion and rewritten in the form
2 3 2 3
Iix Jix Kix tix 1 {Dviz Dviy Ddix
6 Kiy tiy 7 6 {Dvix Ddiy 7
i
Aa ~6
Iiy Jiy 7~6 Dviz 1 7: (24)
4 Iiz Jiz Kiz tiz 5 4 {Dviy Dvix 1 Ddiz 5
0 0 0 1 0 0 0 1
Applying the assumption a A0 ~i A0 zdi A0 ~i Aa -1i A0 , it can be shown that

2 3
Iix zDIix Jix zDJix Kix zDKix tix zDtix
6 I zDI 7
a 6 iy iy Jiy zDJiy Kiy zDKiy tiy zDtiy 7
A0 ~6 7
4 Iiz zDIiz Jiz zDJiz Kiz zDKiz tiz zDtiz 5
0 0 0 1
2 32 3 (25)
1 Dviz Dviy Ddix Iix Jix Kix tix
6 Dv 1 76
Dvix Ddiy 76 Iiy Jiy Kiy tiy 7
6 iz 7
~6 76 7:
4 Dviy Dvix 1 Ddiz 54 Iiz Jiz Kiz tiz 5
0 0 0 1 0 0 0 1
From Eq. (25), it can be deduced that
8 8
>
< DI ix ~I iy Dv iz {I iz Dv iy < DJix ~Jiy Dviz {Jiz Dviy
>
DIiy ~{Iix Dviz zIiz Dvix , DJiy ~{Jix Dviz zJiz Dvix ,
>
: >
:
DIiz ~Iix Dviy {Iiy Dvix DJiz ~Jix Dviy {Jiy Dvix
8 8
< DKix ~Kiy Dviz {Kiz Dviy
> < Dtix ~tiy Dviz {tiz Dviy {Ddix
>
DKiy ~{Kix Dviz zKiz Dvix , Dtiy ~{tix Dviz ztiz Dvix {Ddiy :
>
: >
:
DKiz ~Kix Dviy {Kiy Dvix Dtiz ~tix Dviy {tiy Dvix {Ddiz
Furthermore, differentiating Eqs. (7), (19), and (20), it can be shown that the differential
changes in the incident point position DPi , DVi , and vector DVi are respectively given by

DPi{1
DPi ~DPi{1 zli DVi{1 zVi{1 Dli ~MPi zMP ½Dei , (26)
DVi{1 i

DPi{1
DVi ~MVi zMV ½Dei , (27)
DVi{1 i
and

DPi{1
DVi ~MVi zMV ½Dei : (28)
DVi{1 i
Combining Eqs. (26), (27), and (28), the differential changes in DPi and the refracted
(reflected) ray unit directional vectors DVi (DVi ) can be derived as

DPi DPi{1
~Mi zMi ½ei , (29)
DVi DVi{1
T
where ½ei ~½ Ddix Ddiy Ddiz Dvix Dviy Dviz and the forms of Mi and Mi are presented
in the APPENDIX. The corresponding light path error induced at the (n – 1)-th boundary
surface can then be determined from

DPn{1 DPn{2
~Mn{1 ½en{1 zMn{1
DVn{1 DVn{2

DPn{3
~Mn{1 ½en{1 zMn{1 Mn{2 ½en{2 zMn{1 Mn{2
DVn{3 (30)
~M ½en{1 zMn{1 M ½en{2 zMn{1 Mn{2 M ½en{3 z::::::::::::
n{1 n{2 n{3
zMn{1 Mn{2 Mn{3 Mn{4 :::::M2 M1 ½e1
~Mn{1 ½en{1 zMn{2 ½en{2 zMn{3 ½en{3 z::::zM2 ½e2 zM1 ½e1 :
In Eq. (30), Mi (i~1*n{1) is an error analysis matrix of the i-th boundary surface ri that
can be used to analyze the variation of the exit ray of the optical system. Moreover, Mi is
combined the ray path errors at the i-th boundary surface (i.e. three translational errors and
three rotational errors) with the differential changes induced in the reflected/refracted ray unit
directional vector and incident point by differential changes in the light source and unit
directional vector of the incident ray, i.e.
2 3
LPix LPix LPix LPix LPix LPix
6 Ldix Ldiy Ldiz Lvix Lviy Lviz 7
6 7
6 LPiy LPiy LPiy LPiy LPiy LPiy 7
6 7
6 Ld 7
6 ix Ldiy Ldiz Lvix Lviy Lviz 7
6 LP 7
6 iz LPiz LPiz LPiz LPiz LPiz 7
6 7
Mi ~ 6 7: (31)
6 LVix LVix LVix LVix LVix LVix 7
6 7
6 7
6 LViy LViy LViy LViy LViy LViy 7
6 7
6 7
4 LViz LViz LViz LViz LViz LViz 5
Ldix Ldiy Ldiz Lvix Lviy Lviz
4. SIMULATION OF REFRACTIVE RAY PATH ERRORS AT CYLINDRICAL LENS

This section demonstrates the validity of the proposed error analysis methodology using
the case of optical element with refractive cylindrical boundary surface for illustration pur-
poses. Error analysis using the proposed methodology yields a direct and rapid analytical
expression of the errors of a light ray’s path as it passes through optical elements with
cylindrical boundary surfaces. In the current analysis, it is assumed that the light source
T
is located at P0 ~½ 0 P0y 0 1 . The unit directional vector of the incoming ray
V0 ~½ Cb0 Ca0 Cb0 Sa0 Sb0 0 T (a0 and b0 are the polar angles of the incoming ray) is
refracted by the cylindrical boundary surface. The polar angles of the incoming ray are
a0 ~b0 ~0 and a0 ~0o , b0 ~15o . The radius of the cylindrical boundary surface is R1 ~22:54mm
Fig. 4. Refraction at cylindrical boundary.
and the refractive index ratios are N1 ~1:5, N1 ~2:0, and N1 ~2:5 (see Fig. 4). The relative
configuration of (xyz)0 with respect to cylindrical boundary surface frame (xyz)1 is given by
2 3
1 0 0 0
60 1 0 0 7
1
A0 ~6
4 0 0 1 -30 5:
7 (32)
0 0 0 1
In general, the translational error Ddiz and the rotational error Dviz do not influence the
LPi LVi LPi LVi
position of the exit ray, and hence the variations ( ) and ( ) are equal to zero.
Ldiz Ldiz Lviz Lviz
Then, the error analysis matrix Mi is given by
2 3
LPix LPix LPix LPix
6 Ldix Ldiy 0 Lvix Lviy 0 7
6 7
6 LPiy LPiy LPiy LPiy 7
6 0 07
6 Ld 7
6 ix Ldiy Lvix Lviy 7
6 LPiz LPiz LPiz LPiz 7
6 0 0 7
6 7
6 Ld Ld Lv Lv 7
Mi ~6 ix iy ix iy
7: (33)
6 LVix LVix LVix LVix 7
6 0 0 7
6 Ldix Ldiy Lvix Lviy 7
6 7
6 LViy LViy LViy LViy 7
6 0 07
6 Ldix Ldiy Lvix Lviy 7
6 7
4 LViz LViz LViz LViz 5
0 0
Ldix Ldix Lvix Lviy

Figure 5 illustrates the variations in LP1x LV0y and LP1y LV0y with the light source position
P0y at
the refractive index ratio N1 ~1:5. In Fig. 5, it can be observed that the variation in
LP1x LV0y is antisymmetric with regard to P0y , while the variation in LP1y LV0y is symmetric
with regard to P0y . Moreover, total reflection occurs at a small value of P0yfor the incoming ray
o o T
V0 ~½ cos15 0 sin15 0 . Figure 6 illustrates the variations in LV1x Ld1y and LV1y Ld1y
with the light source position P0y at the refractive index ratio N1 ~1:5. From figure 6, it can be
observed that the variation in LV1x Ld1y is antisymmetric with regard to P0y , while the variation
in LV1y Ld1y is symmetric with regard to P0y .
Figure 7 illustrates the variations in LV1y Lv1x and LV1y Lv1y with P0y as a function of the
refractive index ratio N1for V0 ~½ cos15o 0 sin15o 0 T . From Fig. 7, it can be observed
that the variation in LV1y Lv1x is symmetric with regard to P0y , while the variation in LV1y Lv1y
is antisymmetric with respect to P0y . In Fig. 7, it can be observed that as the value of N1
increases, the total reflection occurs at smaller values of P0y . Figure 8 illustrates the variation in
LV1z =Lv1x with P0y as a function of N1 for V0 ~½ cos15o 0 sin15o 0 T . From Fig. 8, it can
be observed that the variation in LV1z =Lv1x is antisymmetric with respect to P0y .
In general, Figs. 5,8 demonstrate that the sensitivity and error analysis of the boundary
surface frame result in significant deviations in the orientation of the light ray refracted by the
optical element with cylindrical surface. If the sensitivity and error analysis matrix of boundary
surface frame (xyz)i are known, we can obtain the variation in the path of the exit light ray of an
optical system that occurs due to differential changes in the six-degrees of freedom (three
translational and three rotational) of the boundary surface frame (xyz)i by using the continuous

Fig. 5. Variations in LP1x LV0y and LP1y LV0y with P0y for N1 ~1:5.

Fig. 6. Variations in LV1x Ld1y and LV1y Ld1y with P0y for N1 ~1:5.

Fig. 7. Variations in LV1y Lv1x and LV1y Lv1y with P0y as function of N1 .
Fig. 8. Variation in LV1z =Lv1x with P0y as function of N1 .
product of the sensitivity and error analysis matrices. The present results based on the analysis
method presented above demonstrate the importance of performing an error analysis and
subsequently applying the analysis method to the design process of the optical system.
5. CONCLUSION
In this paper, the author uses a homogeneous transformation matrix to define the position
and orientation of a local coordinate frame for establishing each boundary surface and for
tracing the light ray as it passes through a cylindrical boundary surface in accordance with
Snell’s law. Furthermore, the results of this paper present a direct and rapid analytical
expression for the errors of a ray’s light path as the ray passes through optical elements with
cylindrical boundary surfaces. The proposed methodology considers two fundamental sources
of error, namely translational and rotational errors at each boundary surface and the
differential changes in the incident point position and unit directional vector of the refracted/
reflected ray resulting from differential changes in the position and unit directional vector of the
light source.
ACKNOWLEDGMENTS
The financial support provided to this study by the National Science Council of Taiwan
under grant NSC96-2221-E-269-003 is gratefully acknowledged.
REFERENCES
1. Conrady, A.E., Applied Optics and Optical Design, Part I, Oxford Press (1929), reprinted by
Dover Publications, Inc., New York, 1957.
2. Herzberger, M., Modern Geometrical Optics, New York, Interscience Publishers, 1958.
3. Hecht, E., Optics, 3rd Edition, Addison-Wesley Publishing Company, 1998.
4. Lin, P.D., ‘‘Analysis and Modeling of Optical Element and System,’’ J. Eng. Ind. Trans. ASME,
Vol. 116, pp. 101–107, 1994.
5. Forbes, G.W., ‘‘Optical system assessment for design: numerical ray tracing in the Gaussian
pupil,’’ J. Opt. Soc. Am., Vol. A 5, pp. 1943–1956, 1988.
6. Forbes, G.W., ‘‘Accuracy doubling in the determination of the final ray configurations,’’ J. Opt.
Soc. Am., Vol. A 6, pp. 1776–1783, 1989.
7. Stone, B.D., ‘‘Determination of initial ray configurations for asymmetric systems,’’ J. Opt. Soc.
Am., Vol. A 14, pp. 3415–3429, 1997.
8. Lin, P.D. and Tsai, C.Y., ‘‘First-order gradients of skew rays of axis-symmetrical optical
systems,’’ J. Opt. Soc. Am., Vol. A 24, pp. 776–784, 2007.
9. Liao, T.T. and Lin P.D., ‘‘Analysis of Optical Elements with Flat Boundary Surfaces,’’ Appl.
Opt., Vol. 42, pp. 1191–1202, 2003.
10. Lin, P.D. and Lu, C.H., ‘‘Analysis and Design of Optical System by Use of Sensitivity Aanlysis
of Skew Ray Tracing,’’ App. Opt., Vol. 43, pp. 796–807, 2004.
11. Lin, P.D. and Liao, T.T., ‘‘A New Model of Binocular Stereo Coordinate Measurement System
Based on Skew Ray Tracing,’’ J. Dyn. Syst. Meas. Control. Trans. ASME, Vol. 126, pp. 102–
114, 2004.
12. Tsai, C.Y. and Lin, P.D., ‘‘Prism Design Based on Changes in Image Orientation,’’ App. Opt.,
Vol. 45, pp. 3951–2959, 2006.
13. Tlusty, J., ‘‘Techniques for Testing Accuracy of NC-Machine Tools,’’ 12th MDTR-Conference,
1972.
14. Ferreira, P.M. and Liu, R.C., ‘‘An Analytical Quadratic Model for the Geometric Error of a
Machine Tool,’’ J. Manuf. Syst., Vol. 5, pp. 51–63, 1986.
15. Schultshik, R., ‘‘The Components of Volumetric Accuracy,’’ Ann. CIRP., Vol. 25, pp. 223–228,
1977.
16. Mahajan, V.N., Optical Imaging and Aberrations, Washington, SPIE Bellingham, 1998.
17. Paul, R.P., Robot Manipulators-Mathematics, Programming Control, Cambridge, MA, MIT
press, 1982.
APPENDIX
h i
DPi{1
DHi ~ Hi,1 Hi,2 Hi,3 Hi,4 Hi,5 Hi,6 z Hi,1 Hi,2 Hi,3 Hi,4 Hi,5 Hi,6 ½Dei : (A1)
DVi{1
Hi,1 ~Hi,2 ~Hi,3 ~0 Hi,4 ~2(Iix Gix zIiy Giy ) Hi,5 ~2(Jix Gix zJiy Giy ) Hi,6 ~2(Kix Gix zKiy Giy )
Hi,1 ~Hi,2 ~Hi,3 ~Hi,6 ~0 Hi,4 ~2Giy Giz Hi,5 ~{2Gix Giz
h i
DPi{1
DDi ~ Di,1 Di,2 Di,3 Di,4 Di,5 Di,6 z Di,1 Di,2 Di,3 Di,4 Di,5 Di,6 ½Dei (A2)
DVi{1
Di,1 ~Iix Gix zIiy Giy Di,2 ~Jix Gix zJiy Giy Di,3 ~Kix Gix zKiy Giy
Di,4 ~Iix (Fix ztix )zIiy (Fiy ztiy ) Di,5 ~Jix (Fix ztix )zJiy (Fiy ztiy )
Di,6 ~Kix (Fix ztix )zKiy (Fiy ztiy ) Di,1 ~{Gix Di,2 ~{Giy
Di,3 ~0 Di,6 ~0 Di,4 ~Fiy Giz zFiz Giy ztiz Giy ztiy Giz
Di,5 ~{(Fix Giz zFiz Gix ztiz Gix ztix Giz )
h i
DPi{1
DEi ~ Ei,1 Ei,2 Ei,3 Ei,4 Ei,5 Ei,6 z Ei,1 Ei,2 Ei,3 Ei,4 Ei,5 Ei,6 ½Dei (A3)
DVi{1
Ei,1 ~2Iix (Fix ztix )z2Iiy (Fiy ztiy ) Ei,2 ~2Jix (Fix ztix )z2Jiy (Fiy ztiy )
Ei,3 ~2Kix (Fix ztix )z2Kiy (Fiy ztiy ) Ei,4 ~Ei,5 ~Ei,6 ~0 Ei,1 ~{2(Fix ztix )
Ei,2 ~{2(Fiy ztiy ) Ei,3 ~0 Ei,6 ~0
Ei,4 ~2(Fiy Fiz zFiy tiz zFiz tiy ztiy tiz ) Ei,5 ~{2(Fix Fiz zFix tiz zFiz tix ztix tiz )
h i
DPi{1
Dli ~ li,1 li,2 li,3 li,4 li,5 li,6 z li,1 li,2 li,3 li,4 li,5 li,6 ½Dei (A4)
DVi{1

1 1 2 {1 Hi,j h 1
i
li,j ~ {Di,j + (Di {Hi Ei ) (2Di Di,j {Hi,j Ei {Hi Ei,j ) { 2 {Di +(D2i {Hi :Ei )2
: 2 : : :
Hi 2 Hi
(j~1*6)
H h i
1 1 {1 i,j 1
li,j ~ {Di,j + (D2i {Hi :Ei ) 2 (2Di :Di,j {Hi,j :Ei {Hi :Ei,j ) { 2 {Di +(D2i {Hi :Ei )2
Hi 2 Hi
(j~1*6)
h i
DPi{1
Dsi ~ si,1 si,2 si,3 si,4 si,5 si,6 z si,1 si,2 si,3 si,4 si,5 si,6 ½Dei (A5)
DVi{1
si,1 ~Iix zGix li,1 si,2 ~Jix zGix li,2 si,3 ~Kix zGix li,3 si,4 ~Iix li zGix li,4
si,5 ~Jix li zGix li,5 si,6 ~Kix li zGix li,6 si,1 ~{1zGix li,1 si,2 ~Gix li,2 si,3 ~Gix li,3
si,4 ~Gix li,4 si,5 ~{Fiz {li Giz {tiz zGix li,5 si,6 ~Fiy zli Giy ztiy zGix li,6
h i h i
DP
Dri ~ ri,1 ri,2 ri,3 ri,4 ri,5 ri,6 i{1
z ri,1 ri,2 ri,3 ri,4 ri,5 ri,6 ½Dei (A6)
DVi{1
ri,1 ~Iiy zGiy li,1 ri,2 ~Jiy zGiy li,2 ri,3 ~Kiy zGiy li,3 ri,4 ~Iiy li zGiy li,4
ri,5 ~Jiy li zGiy li,5 ri,6 ~Kiy li zGiy li,6 r ~Giy li,1 r ~Giy li,2 {1 r ~Giy li,3
i,1 i,2 i,3
r ~Giy li,4 z(Fiz ztiz )zli Giz r ~Giy li,5 r ~Giy li,6 {(Fix ztix ){li Gix
i,4 i,5 i,6
h i
DPi{1
Dai ~ ai,1 ai,2 ai,3 ai,4 ai,5 ai,6 z ai,1 ai,2 ai,3 ai,4 ai,5 ai,6 ai,7 ½Dei (A7)
DVi{1
si ri,j {ri si,j s i r {ri si,j
i,j
ai,j ~ (j~1*6) ai,j ~ (j~1*6)
s2i zr2i s2i zr2i
2 3 2 3
Dnix nix,1 nix,2 nix,3 nix,4 nix,5 nix,6
6 7 6n 7 DPi{1
4 Dniy 5~4 iy,1 niy,2 niy,3 niy,4 niy,5 niy,6 5
DVi{1
Dniz niz,1 niz,2 niz,3 niz,4 niz,5 niz,6
2 3 (A8)
nix,1 nix,2 nix,3 nix,4 nix,5 nix,6
6 7
z4 niy,1 niy,2 niy,3 niy,4 niy,5 niy,6 5½Dei
niz,1 niz,2 niz,3 niz,4 niz,5 niz,6
nix,j ~si (Iiy Cai {Iix Sai )ai,j niy,j ~si (Jiy Cai {Jix Sai )ai,j niz,j ~si (Kiy Cai {Kix Sai )ai,j (j~1*6)
nix,j ~si (Iiy Cai {Iix Sai )ai,j niy,j ~si (Jiy Cai {Jix Sai )ai,j niz,j ~si (Kiy Cai {Kix Sai )ai,j (j~1*3)
nix,4 ~si (Iiy Cai {Iix Sai )ai,4 zIiz Sai nix,5 ~si (Iiy Cai {Iix Sai )ai,5 {Iiz Cai
nix,6 ~si (Iiy Cai {Iix Sai )ai,6 zIiy Cai {Iix Sai niy,4 ~si (Jiy Cai {Jix Sai )ai,4 zJiz Sai
niy,5 ~si (Jiy Cai {Jix Sai )ai,5 {Jiz Cai niy,6 ~si (Jiy Cai {Jix Sai )ai,6 zJiy Cai {Jix Sai
niz,4 ~si (Kiy Cai {Kix Sai )ai,4 zKiz Sai niz,5 ~si (Kiy Cai {Kix Sai )ai,5 {Kiz Cai
niy,6 ~si (Jiy Cai {Jix Sai )ai,6 zJiy Cai {Jix Sai

DPi{1
DChi ~ Chi,1 Chi,2 Chi,3 Chi,4 Chi,5 Chi,6
DVi{1 (A9)
h i
z Chi,1 Chi,2 Chi,3 Chi,4 Chi,5 Chi,6 ½Dei
Chi,j ~{si (Giy Cai {Gix Sai )ai,j Chi,j ~{si (Giy Cai {Gix Sai )ai,j (j~1*3)
Chi,4 ~{si ½(Giy Cai {Gix Sai )ai,4 zIix Cai zIiy Sai
Chi,5 ~{si ½(Giy Cai {Gix Sai )ai,5 zJix Cai zJiy Sai
Chi,6 ~{si ½(Giy Cai {Gix Sai )ai,4 zKix Cai zKiy Sai Chi,4 ~{si ½(Giy Cai {Gix Sai )ai,4 zGiz Sai
Chi,5 ~{si ½(Giy Cai {Gix Sai )ai,5 {Giz Cai Chi,6 ~{si ½(Giy Cai {Gix Sai )ai,6 zGiy Cai {Gix Sai
The differential changes of a light ray refracted (reflected) by cylindrical boundary surface has
the form

DPi DPi{1
~Mi ½ei zMi Mi ~½mijk Mi ~½mijk (j~1*6, k~1*6) (A10)
DVi DVi{1

DPi DPi{1
~Mi ½ei zMi Mi ~½mijk Mi ~½mijk (j~1*6, k~1*6) (A11)
DVi DVi{1
For the case of a refracted light ray:
mi1k ~d1k zli d4k zVi{1x li,k mi2k ~d2k zli d5k zVi{1y li,k mi3k ~d3k zli d6k zVi{1z li,k
0 1
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
B {nix N2i Chi C
mi4k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi nix AChi,k zNi z({ 1{N2i z(Ni Chi )2 zNi Chi )nix,k
1{N2i z(Ni Chi )2
0 1
B {niy N2i Chi C
mi5k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi niy AChi,k zNi z({ 1{N2i z(Ni Chi )2 zNi Chi )niy,k
1{N2i z(Ni Chi )2
0 1
B {niz N2i Chi C
mi6k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi niz AChi,k zNi z({ 1{N2i z(Ni Chi )2 zNi Chi )niz,k
2 2
1{Ni z(Ni Chi )
mi1k ~Vi{1x li,k mi2k ~Vi{1y li,k mi3k ~Vi{1z li,k

0 1
2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
B {nix Ni Chi C
mi4k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi nix AChi,k {( 1{N2i z(Ni Chi )2 {Ni Chi )nix,k
1{N2i z(Ni Chi )2
0 1
B {niy N2i Chi C
mi5k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi niy AChi,k {( 1{N2i z(Ni Chi )2 {Ni Chi )niy,k
1{N2i z(Ni Chi )2
0 1
B {niz N2i Chi C
mi6k ~@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi zNi niz AChi,k {( 1{N2i z(Ni Chi )2 {Ni Chi )niz,k
1{N2i z(Ni Chi )2
For the case of a reflected light ray:
mi1k ~d1k zli d4k zVi{1x li,k mi2k ~d2k zli d5k zVi{1y li,k mi3k ~d3k zli d6k zVi{1z li,k
mi4k ~d4k z2nix Chi,k z2Chi nix,k mi5k ~d5k z2niy Chi,k z2Chi niy,k mi6k ~d6k z2niz Chi,k z2Chi nix,k
mi1k ~Vi{1x li,k mi2k ~Vi{1y li,k mi3k ~Vi{1z li,k mi4k ~2nix Chi,k z2Chi nix,k
mi5k ~2niy Chi,k z2Chi niy,k mi6k ~2niz Chi,k z2Chi nix,k
NOMENCLATURE
ri i-th boundary surface with unit normal ni .
k
Ag configuration of frame (xyz)g with respect to frame (xyz)k .
(xyz)0 world coordinate frame.
(xyz)i ideal coordinate frame embedded in i-th boundary surface.
(xyz)a actual coordinate frame embedded in i-th boundary surface.
Pi incident point on i-th boundary surface.
Vi unit directional vector of light ray following reflection/refraction at ri .
hi and hi incident angle and refracted angle, respectively.
Ni Ni ~ji{1 =ji , where ji is refractive index of medium i with respect to
vacuum.
DPi differential change in incident point position Pi .
DVi differential change in unit directional vector Vi .
Mi 6|6 sensitivity matrix defined by ½ DPi DVi T ~Mi ½ DPi{1 DVi{1 T
when a light ray passes through the i-th boundary surface ri .
Mi 6|6 error matrix defined by
T
½ DPi DVi T ~Mi ½ Ddix Ddiy Ddiz Dvix Dviy Dviz , when light
ray passes through the i-th boundary surface ri .
Mi 6|6 error analysis matrix based on coordinate frame (xyz)i of the i-th
boundary surface.

Skew Ray Tracing and Error Analysis of Optical Lens With Cylindrical Boundary Surface - Liao Te Ten, Chen Shih Hung, Chen Kuo Ying, Chen Chun Ta

Uploaded by

Copyright:

Available Formats

Skew Ray Tracing and Error Analysis of Optical Lens With Cylindrical Boundary Surface - Liao Te Ten, Chen Shih Hung, Chen Kuo Ying, Chen Chun Ta

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Skew Ray Tracing and Error Analysis of Optical Lens With Cylindrical Boundary Surface - Liao Te Ten, Chen Shih Hung, Chen Kuo Ying, Chen Chun Ta

Uploaded by

Copyright:

Available Formats

SKEW RAY TRACING AND ERROR ANALYSIS OF OPTICAL LENS WITH

CYLINDRICAL BOUNDARY SURFACE

Received February 2008, Accepted January 2009

LANCER DE RAYONS OBLIQUES ET ANALYSE DE LENTILLES OPTIQUES À

2. SKEW RAY TRACING FOR CYLINDRICAL LENS

P0i ~½ Pi{1x zVi{1x l Pi{1y zVi{1y l Pi{1z zVi{1z l 1 T , (6)

Pi ~½ Pi{1x zVi{1x li Pi{1y zVi{1y li Pi{1z zVi{1z li 1 T (7)

Fig. 2. Physical meaning of configuration matrix i A0 .

Hi ~G2ix zG2iy , (9)

Di ~Gix (Fix ztix )zGiy (Fiy ztiy ), (10)

Ei ~(Fix ztix )2 z(Fiy ztiy )2 {R2i : (11)

hi ~Fiz zli Giz ztiz , (13)

where si ~Fix zli Gix ztix , ri ~Fiy zli Giy ztiy .

Chi ~{VTi{1 :ni ~si (Gix Cai zGiy Sai ): (14)

mi ~½ mix miy miz 0 T ~ni |Vi{1=jni |Vi{1 j~ni |Vi{1=Shi : (16)

3. ERROR ANALYSIS FOR CYLINDRICAL LENS

Applying the assumption a A0 ~i A0 zdi A0 ~i Aa -1i A0 , it can be shown that

4. SIMULATION OF REFRACTIVE RAY PATH ERRORS AT CYLINDRICAL LENS

Fig. 4. Refraction at cylindrical boundary.

Fig. 8. Variation in LV1z =Lv1x with P0y as function of N1 .

For the case of a refracted light ray:

mi1k ~Vi{1x li,k mi2k ~Vi{1y li,k mi3k ~Vi{1z li,k

For the case of a reflected light ray:

You might also like