Looking Back: A Lens Designer Remembers

Nearly 57 years ago, in 1962, I began work as an optical engineer with General Electric (GE) in Scranton, Pa. As an engineering technician, I worked with GE’s senior optical engineer, Don Kienholz (Figure 1). The computer system we used at the time included a specialty software package that had lens design and optimization capabilities. The software had been developed specifically for IBM by Gordon Spencer of Scientific Calculation Inc. in Rochester, N.Y.

Figure 1. Don Kienholz (right) working with the author at GE’s rented IBM 1620 computer in 1963. Courtesy of Bruce H. Walker.

After a year of training, Don assigned my first design. It was an IR objective lens that would be used to create a high-quality image on the screen of an IR video sensor. This system would function at the specific wavelength of 1.05 µm (laser energy). The specifications of the lens included an EFL (effective focal length) of 10.25 in. (260 mm), a field of view of 5.5°, and a lens aperture of f/3.0. The lens needed to fit into a package that was 11 in. long and about 5 in. in diameter.

When I met with Don and Robert Hopkins (of the University of Rochester), it was decided that the lens would be a basic Petzval lens form, adjusted to make the lens’ overall length nearly equal to the 260-mm EFL of the lens. Petzval lenses were developed in 1840 in Vienna by the German-Hungarian mathematics professor Josef Maximillián Petzval. They were the first portrait objective lenses in the history of photography.

My design took nearly three months to complete. I am happy to say I still have the spec and image quality data (Figure 2). In its four-quadrant layout, the data included the following: a table of all lens numeric data that enabled the final detail lens drawings to be generated (upper left); a drawing of the complete lens and the beams of light for the on-axis points and the maximum field of view, which was 2.75° (upper right); the modulation transfer functions (MTF) for the on-axis and three off-axis points (lower left); and the lens image quality performance at 50 cycles/mm. This indicated the image quality would range from 0.58 modulation (contrast) on-axis to 0.38 modulation (contrast) at the edge of the image. The distortion of this lens was corrected essentially to zero over the full field of view, which was 5.5° square (lower right).

Although Figure 2 shows computer printouts of the design, I feel it’s interesting to point out that in 1964, when this design was done, only the numeric data was available from the IBM 1620. It was left to the designer to hand draw the complete lens, which would take approximately four hours with the use of a drafting board and drafting tools. It would take three additional hours for the MTF plotting, again using the aforementioned tools as well as graph paper. With today’s PC and software, those two quadrant images are typically completed in seconds!

Figure 2. The author’s first lens design, with lens data, 1964. Courtesy of Bruce H. Walker.

My career with GE lasted about eight years, from 1963 until 1971. This time included a relocation of my then-growing family — along with the other GE optical engineering department families — to Syracuse, N.Y. During this time, the focus of my optical lens design engineering would include head-up displays, gunsights for helicopters, star sight optics, objective lenses for medical x-ray systems, and numerous others.

In 1968, I had my first writing piece published. In the November/December issue of Optical Spectra (now Photonics Spectra), you can find my featured article “Specifications of Optical Performance.” My interest in writing would lead to a nurturing relationship with Laurin Publishing. In particular, Teddi Laurin’s encouragement would lead to my writing more than 40 articles, two textbooks, and a history of my eventual employer, Kollmorgen Corp., in book form. In 1971, GE decided to eliminate the lens design subgroup of the optical engineering department. I was encouraged by GE to stay on in a different optical capacity, but my heart lay in lens design. So again our growing family was on the move.

For the next 20 years, I did optical engineering and lens design work with Kollmorgen in Northampton, Mass. This work involved generating the lens designs of military gunsights, submarine periscopes, and head-up optics for the A-10 attack aircraft, and numerous other projects. During this time, I would grow to become a mentor to up-and-coming optical engineers, as well as many young engineers in other departments. After 10 years with the company, I became manager of the Optical Engineering Group.

Then I moved into a career as an optical engineering consultant, which has lasted for the past 25 years. I continued my established lens-design work while taking on work in new major optics fields, including visual optics and the human eye, specification and design of the eyepiece, and 3D stereo optical systems.

Over five decades in optical engineering, I — like all who had been in the field — watched as computers and optical design software grew in an extraordinary fashion. I started out on customized IBM software, then used Automatic Correction of Centered Optical Systems (ACCOS) — an improved software for complex designs — and later Code III to Code V, which (at the time) was typically available through leasing because of the software’s expense and sophisticated computing requirements.

My work as a consulting engineer followed a trajectory somewhat parallel to the development of the PC, along with affordable software packages for lens design and evaluation. As a consulting engineer, I was fortunate to begin working with the OSLO (Optics Software for Layout and Organization) lens design program developed by Doug Sinclair of Sinclair Optics.

In 2015, nearing the end of my optical engineering career, I stumbled upon the data for my first lens design. At this time, I was working with the premium version of the OSLO 6.5 software package that was provided and supported by Lambda Research Corp. of Littleton, Mass. My goal with OSLO 6.5 was (as it often was) to create a new design that would be improved in all regards: image quality, simplicity, and cost. In 2015, I spent a total of four hours to generate a new design, compared with the two to three months it took in 1964.

Figure 3 shows the results of that new design. Noteworthy is its simplification because of the elimination of the field-flattening lens from the original design. The image quality of the new design includes MTF, which has been improved to be consistent to 0.71 contrast at 50 cycles/mm for all points within the full field of view. Similar to my first design, the new design is held at essentially zero distortion over the full field of view.

Figure 3. A new design, with lens data, 2015. Courtesy of Bruce H. Walker.

Over the course of 50 years, I was lucky to have experienced the technological advancements in computer hardware, while at the same time (with the help of mentors) crafting my skills as a young lens designer. Looking back, I realize that it was an unusual synergy; neither experience could have occurred without the other.

Meet the author

Bruce H. Walker has worked in the fields of optical engineering and lens design since 1960, initially with GE and later with Kollmorgen Corp. Since 1991, he’s been an independent consultant specializing in solving optical engineering problems and creating lens designs.