Testing and centering of lenses by means of a Hartmann test with four holes Daniel Malacara, FELLOW SPIE Zacarias Malacara Centro de Investigaciones en Optica Apartado Postal 948 LeOn, Mexico Abstract. The Hartmann test is frequently performed with a screen with many holes to test the quality of a mirror or the aberrations of a lens. An application of the simplest form of this test with only four holes is described. This test cannot detect spherical aberration, but it may be useful to align optical systems, to detect and measure focus errors, or to detect lens decenterings and coma. Subject terms: optical testing; geometrical optics; Hartmann test. Optical Engineering 3 1(7), 1551-1555 (July 1992). 1 Introduction In this paper we describe a modified and simplified version of the Hartmann test. The Hartmann test1 is frequently performed with a screen with many holes, as shown in Fig. 1, to test the shape of optical surfaces.2 The simplest Hartmann screen that may be devised is one with only four holes in a circle, separated by an angle of 90 deg. This test is so simple that it cannot detect spherical aberration, but it may be useful to align optical systems, to detect and measure focus errors, or to detect lens decenterings and coma. 2 Theory Let us assume that the wavefront deformations in a lens system, relative to a spherical wavefront, can be represented by: W=P(x cosa+y sina)+0.5S(x2+y2) +0.5C(x sin—y cos)2 +B(x cosl3 +y sinl3)(x2 +y2) (1) where the decentering of the lens or its prismatic error is P, which produces a wavefront tilt, measured in radians. Paper 15081 received Aug. 29, 1991; revised manuscript received Nov. 29, 1991; accepted for publication Nov. 30, 1991. 1992 Society of Photo-Optical Instrumentation Engineers. 0091-3286/92/$2.00. This decentering is in the direction a. In other words, the wavefront is tilted about an axis in the perpendicular direction to a. The focus error S is measured as a spherical power with respect to the reference wavefront. It is proportional to the curvature of the wavefront, measured with respect to this reference, and defined as the inverse of the radius of curvature of this wavefront. The astigmatism is C, which may be produced by a decentering of a lens or by a toroidal surface in the system. It is the power of the toroidal or cylindrical component of the wavefront, which is proportional to its curvature. The curvature along the direction is equal to the focus error or spherical curvature, while the curvature in the perpendicular direction is the sum of this spherical curvature plus the cylindrical curvature. If the optical curvatures are all spherical, this direction is perpendicular or parallel to the decentering direction a of the lens, depending on the sign of the astigmatism. The coma component of the wavefront is B, in units of length — 2 The coma is in the direction 3 , which is the same direction in which the lens is decentered. Coma may appear whenever there is a decentering or prismatic error in the lens. Notice that all decentering and tilts of the lens under test are relative to an optical axis defined by the diaphragm with the four holes. This axis is a line perpendicular to the diaphragm that passes through the center between the holes. OPTICAL ENGINEERING / July 1 992 / Vol. 31 No. 7 / 1551 Downloaded From: http://opticalengineering.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms MALACARA and MALACARA DIAPHRAGM b C LENS UNDER TEST Fig. 1 Hartmann test layout. Fig. 2 Four holes to make Hartmann test. The wavefront deformations W and the transverse abrelated by the following well- errations TA and TA are known expressions: TA = -f () TA (2) and S = 0.1 = TA=_f() TA TA = 0° (3) . Thus , using these relations , the x components of the trans- TA verse aberrations are given by: +2Bx(x cos3+y sin3)+B(x2+y2) cos3 , P sina + Sy — C(x sin —y __7__ 0 = 45' = 45U (4) Fig. 3 Images for a combination offocusing error (S = 0.1 and S = 1.0) 45 deg). and astigmatism (C = 1 with 4 = 0 deg and = and the y components by — TA = S=1 S = 0.1 — = P cosa + Sx + C(x sin — y cos4) sin cos) cos + 2By(x cos3 + y sin3) + B(x2 + y2) sin3 , (5) where f is the distance from the wavefront to the measuring plane for the transverse aberrations . If the four holes on the Hartmann screen are represented by a, b, c, and d, as shown in Fig. 2, we may find: p= 0 TAxa/fP cosa + p(S + C sin2) + 3Bp2 cos3 — TAbIf= P cosa — pC sin cos + Bp2 cos3 — — TA/f= P cosa — p(S + C sin2) + 3Bp2 cos3 — TAdIf= P cosc + pC sin cos + Bp2 cos3 — TAya/f P sina — pG 0° = 0° TA (6) sin cos + Bp2 sin3 TAyb/fP sina+p(S+C cos2)+3Bp2 sin3 — TA/f= P sina + pC sin cos + Bp2 sin3 TAyd/fP sina—p(S+C cos2)+3Bp2 sin3 .. p = 0' 0 = 45° 45' 45' Examples of these transverse aberrations for a mixture of focusing errors and astigmatism are given in Fig . 3 , for a mixture of coma and astigmatism in Fig. 4 and for a mixture Fig. 4 Images for a combination of coma (B = 0.5 with 3 — 0 deg and of focusing error and coma in Fig. 5. 1=45 deg) and astigmatism (C=1 with =0 deg and =45 deg). 1 552 / OPTICAL ENGINEERING / July 1 992 / Vol. 31 No. 7 Downloaded From: http://opticalengineering.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms TESTING AND CENTERING OF LENSES and from these, we obtain: TA TA C = [(ixa iyb) + (ixb + ie Sya)21112 . (9) TA This is the astigmatism power, or cylinder, which has an axis orientation 4 given by: I. S = 0.1 P = 0' 0' TA TA i. S = tan24= •TA xb+ya (10) °xa — Oyb The focus error, also called sometimes spherical power, is given by: s= 1 0.1 p = 45' p = 45' 5Sxa+yb+C . Fig. 5 Images for a combination of focusing error (S=0.1 and S= 1.0) and coma (B=0.5 with 13=0 deg and 3=45 deg). (11) The lens decentering, wavefront tilt or prismatic error, is given by: = p[(3a- — a-yb) + (a-xa — 3a-xb)2]112 If the differences of these aberrations divided by 2pf are represented by S, and the sums divided by 2pf are represented by a-, we find: xa = TAxa — TA = — — C sin2 tana= 2pf TAya — TA = C sin cos yb = TAb—TAd —s—c cos24 a-xa= crxb= a-yb= (7) TAxa + TAxc = — P cos — p 2pf TAxb+TAxd P Cosa 2pf p P Sifla p tan3= xb + yb2SC ya = 2C xb ya — ( . a-yb) + (a-xa a-xb)2]112 (14) a-ya a-yb a-xb a-xa . (15) following term T has to be zero, unless the original assumption that the wavefront can be represented by Eq. (1) is not valid, thus indicating the presence of higher order aberrations or a bad measurement: 3Bp sinl3 T6xbya . (16) From If all the surfaces are spherical and centered along a common axis, the angles a and 3 should be the same and sin cos4 = C sin2 yb = crxa The 0 — C(cos24 sin2) = C cos24 = — 2Bp a-xb COSI3 — 2(P/p) CO5 3a-xb = — a-yb = 2Bp sinl3 a-ya 3a-ya a-yb= —2(P/p) sina xa — 3crxb B cos3 where all these quantities have the units of length these expressions we now easily find: axa+ a-yb which has an orientation given by: 3Bp cosl3 TAya +TA = B sinl3 2pf TAN,, + TAd = — P sina — 2pf 3a-ya B = [(a-ya 2pf 2pf has an orientation given by: The coma aberration is: xb TAxb_TAxdC sin cos ya = which (12) (8) should also be the same, or mutually perpendicular to the first two. If all the surfaces are spherical and the angles a and 3 are not equal and/or the angle is not perpendicular to them, this is because all the surfaces are not centered along a common axis. Obviously the two spherical surfaces in a single lens are always centered along a common axis. If we may assume that there are no higher order aberrations or bad measurements that make T different from zero, we may write: OPTICAL ENGINEERING / July 1 992 / Vol. 31 No. 7 / 1553 Downloaded From: http://opticalengineering.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms MALACARA and MALACARA DIAPHRAGM I PtJtT1DN SENSDIG D(TECTOP LENS UNDER TEST (a) PosrnoN SEPGING JIETECTDR LD (a) UNDER TEST LIGHT SOURCE DIAPHRAGM 4 POSITION SENSING DETECTORS POSITION SENSING DETECTOR (b) (b) LENS MIRROR UNDER TEST UNDER TEST Fig. 6 Measurement of a lens or a mirror with large focal length. Fig. 7 Optical configurations to measure the transverse aberrations: (a) at the focus, with a small position sensing detector and (b) out of focus, with four small position sensor detectors with a television camera or with a photographic plate. xb ya The optical arrangement to take the measurements may be implemented as in Fig. 6(a) or as in Fig. 6(b), with an autocollimation arrangement, to measure concave optical surfaces. In this manner, the radius of curvature may be measured. If the coma and astigmatism are determined, the optical center of an aspheric surface may be located. The and 0yaffyb0xaffxbO then, the astigmatism would have a magnitude: C = {(xa iyb) + 1/2 (17) with an orientation: 2ab tan2= xa yb (18) second lens is introduced to keep the length of the instrument small, even if the focal length of the element under test is large. If a single-point light source is used, as in the classical Hartmann test, a large defocusing has to be introduced to separate the four images. Then, their positions may be measured using a photographic plate, a television camera, or four-position sensing detectors, as in Fig. 7(b). A single small position sensing detector may be used at the focus, as shown in Fig. 7(a), if the four holes are illuminated one at a time to take the four measurements . This and the focusing error or spherical power would be: can also be performed in many ways. Just to give a few 2 ''19I Note that the transverse aberration measurements may be absolute as well as relative. For example, considering Fig. 6(a), the lens after the diaphragm with the four holes may have large aberrations, but then a second lens is inserted in front of the diaphragm. If the lens under test is the one in front, the transverse aberrations to be measured are those representing the difference between those without and with the lens. Then, two measurements are made, one before and one after the lens to be tested is inserted. examples, let us consider the four methods in Fig. 8. Figure 8(a) shows the use of a rotating slotted shutter in front of the diaphragm with four holes. In Fig. 8(b), a laser scanning device is used. It could be made with galvanometers, acoustooptic deflectors, or any other kind of deflector. Four independent light sources with a small circular diaphragm at the focus of a collimating lens are shown in Fig. 8(c). Finally, four independent light sources with four small lenses are shown in Fig. 8(d). 3 Conclusions This simple test is quite useful for examining the alignment accuracy of many systems, or for measuring the coma and 1554 / OPTICAL ENGINEERING / July 1992 / Vol. 31 No. 7 Downloaded From: http://opticalengineering.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms TESTING AND CENTERING OF LENSES Hi ROTATING HUTTER UGHT SOURCE (a) m - (c) (b) FOUR LIGHT SOURCES (d) Fig. 8 Four manners to measure the transverse aberrations when a small position sensing detector is used: (a) with a rotating shutter, (b) with a laser beam scanner, (c) with four light sources and a small circular diaphragm at the focus of a lens, and (d) with four independent light sources and a small lens in front of each. the spherical as well as the cylindrical power of lenses. For example, using this method, a spherometer, a lens measuring instrument, a lens centering machine, or an instrument to locate the vertex of an aspheric surface may be constructed. This test is especially well adapted for use with microprocessors, to make it automatic. References 1. J. Hartmann, "Bemerkungen uber der Bau und die Justirung von Spectrographen," Zf. Instrumentenk. 20, 4 (1900). 2. I. Gosheil, "Hartmann Test," in Optical Shop Testing, Chap. 10, John Wiley, New York (1978). Many other references are included in this chapter. Daniel Malacara obtained his BSc in physics in 1961 from the National University of Mexico and his PhD in optics in 1965 from the University of Rochester, New York. Then, he joined the Institute of Astronomy of the National University of Mexico and in 1972 the Instituto Nacional de Astrofisica, Optica y ElectrOnica. In 1980 the Centro de Investigaciones en Optica was created, and he became its first general director, a po- ——--— — - sition he held from 1 980 to 1 989. He has been the author or editor of several books on optics. The best known is the book Optical Shop Testing. His scientific production includes more than 100 papers published in well-known optics journals. He is a fellow, both of the OSA and SPIE. He is a topical editor for Applied Optics and has been on many international scientific committees. In 1986 he received the National Prize for Technology in Mexico. Zacarlas Malacara obtained a degree in physicsat National University of Mexico in 1 976. He then attended the Institute of Optics at the University of Rochester from 1976 to 1978. After two years at the National Observatory of Mexico, he became an associate researcher at the Center for Research in Optics, where he has been involved in ophthalmic and astronomical optics and optical instrumentation. OPTICAL ENGINEERING / July 1 992 / Vol. 31 No. 7 / 1555 Downloaded From: http://opticalengineering.spiedigitallibrary.org/ on 09/22/2013 Terms of Use: http://spiedl.org/terms