Pioneering the Future with Optical Elements
September 2, 2019
Sony's G Master brand of top-in-class interchangeable lenses is recognized for the superb images it captures by photographers all over the world. The key element is the XA lens, an extreme aspherical lens crafted to nanometer precision. How is it that Sony, a manufacturer that was relatively late to the optical game, can outperform the competition? We asked three engineers who have played key roles in the research and development of this cutting-edge optical element to help us explore this situation as we look back on the last 20 years of aspherical lens development at Sony.
Optical Element Design Dept.,
Core Technology Div. 1,
Digital Imaging Group
Sony Imaging Products
& Solutions Inc.
Opto Design Dept.,
Core Technology Div. 1,
Digital Imaging Group
Sony Imaging Products
& Solutions Inc.
Manufacturing Technology Dept. 2,
Sony Global Manufacturing & Operations Corporation
In the beginning, we were really in the dark
──I imagine you faced a variety of difficulties in the development of the XA lens. Could you tell us how Sony got started developing aspherical lenses?
Hitoshi Nakanishi：The history of Sony's aspherical lens development dates back some 20 years. At first, we set out to develop high-precision, small aspherical lenses that could only be made with Sony's in-house technology, intending to use them in camcorders and digital still cameras. The challenge was that, as the diameter of an aspherical lens increases, the degree of processing difficulty also goes up. So we first had to build some technology for processing small lenses with high precision, and then we gradually stepped up to larger lenses as we established a mass production system.
Daisuke Kuroda：Since camcorders at that time were not yet HD or 4K, and the resolution of the camera's image sensor was low, the performance required for the lens was not very high. However, with the rapid spread of digital cameras and the user demand for high resolution, we had to increase not only the pixel of sensors but also the resolution of lenses. To make that happen, we needed a good aspherical lens that could efficiency correct optical aberration with a single lens element and which could also contribute to compact size and high performance. It was also clear that we wouldn't be able to differentiate our camera products if we used lenses made by another company. So we decided to make the major move to in-house production of aspherical lenses.
Toshihiro Masuda：At that time, however, we didn't have anyone with experience producing aspherical lenses — not just that we didn’t know how to evaluate them, but also that we didn’t even know how to make them. We had to start from scratch and learn about everything from scratch — from the raw materials and the fabrication processes to the facilities needed for mass production. We often felt like we were fumbling around in the dark.
Optical elements shaped by 20 years of R&D
──How long did it take to get to the point that you could mass-produce aspherical lenses?
Masuda：The first mass production run came after three years of development. We called it a hybrid lens since it employed resin aspherical surfaces on spherical glass. To tell you the truth, though, that first run was a failure. The yield was bad and we had to stop production after just two months. We had a problem with a pattern appearing on the surface of the lens. However, trying to solve the problem quickly would have raised other issues that could ruin the product value. We ended up taking about a year to identify all the concerns and then improve the manufacturing equipment and processes to resolve the overall problem.
So, with that experience under our belts, we were finally able to get mass production going, but it took another 10 years for us to transition to molded glass aspherical lenses, which is the current mainstream version. This was because lenses made by melting and molding glass faced big hurdles, not just in terms of technology, but also in terms of the high capital investment required. In fact, it would have been cheaper to buy the lenses from another company. However, management decided that we should expand Sony’s own optical technologies even if it would cost some money. So we continued to do R&D into molded glass lenses while mass producing the hybrid lenses.
Nakanishi：At that time, we held ourselves to a manufacturing standard of staying within submicron deviation of the design value. Technology that could process a lens at that level of precision didn't yet exist anywhere in the world, so we had to develop it ourselves. In order to achieve this, we first developed the technology to process the die with sub-micron precision, and then after many iterations of updating of the equipment to create the formation process, we finally succeeded in creating an advanced aspherical lens (AA lens). From there, we further improved the precision, leading to our XA lens, the extreme aspherical lens which is currently employed in our G Master lenses. All said and done, it took us about 20 years to establish manufacturing technology that could control the surface precision of the lens at the nanometer level. Since it takes this kind of time and money to develop lenses, I doubt the G Master brand would ever have come about had it not been for the foresight and determination of our senior peers.
The close technical collaboration behind the G Master brand
──What role do you each play in the technical development of the XA lens?
Nakanishi：I am in charge of XA lens design. For example, I work on developing the optical glass material, developing new lens shapes, developing the die, and designing the casting process. Most people would not imagine, but there are more than 200 kinds of optical glass materials in the world. Only some of them, however, can be processed into aspherical lenses. So, in order to develop optical elements that do not yet exist, I also have to work on developing materials such as optical glass materials and die materials. I might develop a lens using a material with high hardness close to that of diamond, or a material with very low hardness that is, chemically-speaking, a fragile material. Some lenses may take several years from material development to fabrication and casting process design to completion.
Kuroda：I am in charge of the optical design for G Master and other lenses. I work on creating lens arrangements and group configurations that optimize optical performance. I also use a simulator to discover within the limited size and specifications, for example, the most effective placement for the XA element, to maximize the performance of the lens as a whole. Sometimes I tell Nakanishi-san the kind of shape and refractive index I want, and then he will do the actual process design and technology development to produce a new optical element.
Masuda：My job is to develop the manufacturing technology and equipment for mass production based on the basic process designed by Nakanishi-san. In developing a lens fabrication process, there is a base process, and then I customize it for each purpose. That is, even if the same glass material is molded under the same conditions, it won't necessarily come out as intended if the lens shape is different. So, the recipe for the fabrication process is different for each lens, and a great deal of trial and error is needed to finally achieve stable mass production. If there is a problem, I may have to revise the equipment itself. Since we can't change things so easily once production starts, I always try to discuss things with Nakanishi-san and Kuroda-san early on, in the design phase. I may say, for example, “It’s going to be difficult to achieve mass production with these specs,” and he may respond, “What about this design then?” We work together to increase the precision of the XA lens.
──G Master lenses are gaining recognition among photographers throughout the world for their rendering performance. What advantages do they have over other companies’ products?
Nakanishi：The G Master integrates high resolution and spectacular bokeh at the highest level. Intrinsically, high resolution and beautiful bokeh are mutually exclusive in terms of optics, which makes it very difficult to achieve both. However, the XA lens was the key element that made both possible. The XA lens achieves high surface precision with very little variance from design value and high surface roughness with a very smooth lens surface. In particular, in terms of surface roughness, the XA achieves single-digit nanometer level flatness. Thanks to this property, it successfully eliminates onion-ring bokeh when shooting point light sources, which was a problem for aspherical lenses for a long time. The combination of nanometer level flatness and high surface precision is an advantage that other companies cannot imitate.
Masuda：Onion-ring bokeh is caused by the slightest grinding grooves being left on the lens surface, but in order to eliminate those grinding grooves, it is also necessary to perform grinding and polishing on the nanometer order, which in turn depends on the precision control of the fabrication machinery and the level of skill and knowhow of the machine operator. Aspherical lenses have a complicated lens surface, so it is all the more difficult to do such grinding. We have put a lot of creativity into addressing this issue by combining elements such as tool diameter and rotation pitch in order to reduce grooves as much as possible or to make them in a way that is easy to polish out. In this way, we were able to grind at a high level of precision to eliminate deviations in the lens shape.
Kuroda：When we first started manufacturing XA lenses, we could only make the diameter as big as 40 mm, so the standard and mid-telephoto lens were the mainstream in our lens design. In recent years, manufacturing technology has advanced to the point that we can now make the diameter about 60 mm or more, and this allows us to design wide-angle zoom lenses such as the FE 16-35 mm F2.8 GM. In other words, it is no exaggeration to say that the G Master lineup is determined by the evolution of optical elements.
Using technology to resolve conflicting features
──What are the difficulties and challenges in developing and manufacturing the XA lens?
Masuda：Specifically speaking about the production process for aspherical lenses, first the glass material is heated and pressed to transfer the shape of the die. Then, the pressed glass is annealed, whereby the glass is gradually cooled down and the stress accumulated inside the lens at the time of stamping is released. If this stress is not released well, the lens might break, or the refractive index might be altered, or other problems emerge. After that process, the lens undergoes grinding to form the lens outer shape so that there will be no eccentricity in the optical axis. The next step is the application of an antireflective coating and a black coating to suppress irregular reflection of light from the outer peripheral edge. That is what it takes to make an XM lens.
Nakanishi：Since the glass elements differ from lens to lens, stamping conditions such as molding temperature, pressure, and cooling time are different for each type of lens. There are so many forming parameters to be considered, and there are innumerable combinations, so it takes a very long time to figure out the optimal process conditions. Today, since we have accumulated a lot of testing data, we can now utilize a design simulation, but in the past we had to start with a lot of trial and error, sometimes taking two to three months to optimize a single condition.
Masuda：Glass material repeatedly expands and contracts, so it is pretty difficult to perfectly transfer the shape of the die. For example, even at temperatures of 600°C or higher, a mere 0.1°C difference can lead to an error in the precision of the lens. If the contraction factor fluctuates, you get shape errors, or if the stress remains, you get variance in the refractive index. That is how sensitive it is, so controls need to be very strict. In fact, commercial temperature measurement devices used for lens molding machines cannot meet Sony's requirements. So, sometimes we also need to develop such temperature measurement devices. In the end, we typically have to use quite a bit of creativity across all the processes and equipment to make it all work. Because we now possess the processes, equipment, and knowhow suited to the challenge, we are able to mass-produce the XA lens, which has an overwhelming performance advantage over the competition.
In addition, if precision drops off as we work toward mass production, we need to troubleshoot the cause. However, at the nanometer level, you can't simply look and see what is wrong. Not only that, since aspherical lenses have an asymmetric shape on both sides, the heat transfer and deformation processes are complicated, and it may not be clear why a particular phenomenon is occurring. In such situations, we have to rely on our imagination. Sometimes, you get lucky, and you are able to solve a problem based on prior experience, but more often than not, the phenomenon at hand contradicts everything you have experienced and learned to that point, and no amount of thinking seems to result in a solution. In such cases, you have to be willing to let go of what you thought was common sense, and allow your imagination to take over.
──Are there similar difficulties and challenges in the optical development of G Master lenses?
Kuroda：The difficult part in designing G Master series lenses is to identify design criteria that deliver the combination of high resolution and spectacular bokeh. The key is devising a way to quantify the subjective appreciation of bokeh. Without a way to make a quantifiable evaluation of something, you cannot design for it. So we developed a measurement instrument that could quantify bokeh, and we also made it possible to do an optical simulation of bokeh. This has allowed us to deduce the limit values for the design of each element, thereby making it possible to optimize them for the lens as a whole. This was truly a major breakthrough in designing the G Master. Although it took about three or four years to develop the first simulator, doing so — along with all the data we had accumulated and the visualization of that data — made it possible to finally find a way to manufacture lenses that can deliver both resolution and bokeh.
We didn't stop there in our development of the G Master. We consider ways to make the lens more compact, even though typically the higher the performance requirement, the larger the lens has to be. We consider how we can maintain focus precision, which typically drops off when AF speed is increased. The G Master is, in this regard, a collection of supposedly contradictory features. However, because our basic development philosophy for the G Master is to give up nothing and accept no compromises, we continue to challenge ourselves and do our utmost on every single challenge. Sony's imaging technology cannot evolve without this kind of passion to solve seeming conflicts using technology and our motivation as engineers to create new things.
──With what convictions and goals do you conduct your technology development?
Kuroda：Foremost, we are tasked with the major mission of developing products for customers. Engineers tend to constantly seek out something new, which can sometimes cause us to forget about what customers want. Such a selfishly-developed product is something a customer will see and likely say, "I don't really need that sort of thing." If that happens, then it isn't just sad for our customers, but for our products and ourselves. That's why we always try to think of what will really benefit our customers.
Masuda：The role of the manufacturing is to steadily make and supply products. However, as an engineer, I take pride and am motivated by the fact that we were able to leverage our process to create lenses that stand apart from the competition and then provide them to many customers. Even though the optical element is just one component in the overall product, without it, the supply of those products would come to a halt. Design, assembly, and manufacturing units work together to make our products, and that all builds the G Master brand. As such, I feel a keen sense of responsibility for making sure we succeed.
Creating new technology based on what we already have today
──What kind of human resources do you consider necessary to further advance and develop the technology of optical elements, going forward?
Nakanishi：People who like new things make the best engineers. Right now, we are developing aspherical lenses, but we need people that can effectively use their imagination to take on the challenge of coming up with the next new element. Those kind of people can handle the technology we have now and also think of the future technology that comes next.
Masuda：In our field, nothing is ever the same as it was before, even when it comes to creating a single process. Experience is, of course, important, but we cannot move forward without a mindset that pursues new possibilities. Around here we say, “Of course our current technology has to be competitive, but the really important thing is creating new technology based on the technology we already have.”
Kuroda：The evolution of optical elements is typically continuous, but there may be times when you need something non-continuous. You can only create something that goes beyond common sense by taking an approach that you had never considered before. For example, in the development of the FE 400mm F2.8 GM OSS super telephoto single focus lens, Sony achieved a weight reduction of more than one kilogram while further improving optical performance over that of the competition. In fact, this product was designed using an approach totally different from that of conventional lens design. The idea came from a simple comment uttered by a designer. In other words, to create a breakthrough, you have to be willing to deny even your current self. When people of that mindset come together, it leads to diverse opinions, which in turn leads to evolution. I think this is one of Sony's strengths — that is, the ability to gather diverse opinions and quickly commercialize them.
──Lastly, tell us about your vision for optical element R&D.
Kuroda：Our mission as engineers is to create something that doesn't currently exist in the world. Actually, ten years ago, we often heard people say, What?!? Sony’s making lenses?” We are now finally able to share the G Master with the world, and it is starting to earn broad recognition among photographers. First of all, it is our top priority to do a good job in providing products that our customers want. Then, it is necessary to provide new visual experiences by approaching other business areas with the lens and imaging technology that we have accumulated.
Nakanishi：While there are many areas yet to be explored when it comes to optical elements, we have another development effort already in progress that would revolutionize optical systems. Although I cannot give specifics, the result might not necessarily be in the form of optical elements. The next generation of communication using 5G and evolution via AI will require elements with completely different characteristics. At that time, if we do not have a system in place to deliver what is needed, we will lose our raison d'etre. That is why it is necessary to constantly enhance our sensitivities to the world of the future. I hope that everyone involved in development, design, and mass production will keep working as one to provide new optical elements that can deliver untold new experiences to the world.
Envisioning the Future Created by Sony’s AI
— Artificial Intelligence That Assists Creativity
August 15, 2019