Finding hidden value in high-precision manufacturing

"There is real economic value in high-precision manufacturing," Stack told OLE. "Done properly, it reduces the time-to-market and the value of that is phenomenal. Our expertise enables customers to get their products to market faster and ahead of their competitors. Value-over-time is also important and this approach tends to keep the product viable in the market for longer."

Assembly

High-precision basics
There are distinct differences between precision and high-precision manufacturing. "Precision optics fit together and achieve their desired performance using a 'drop-in' component model," explained Stack. "Although the goal is always to push for drop-in production methods, it is ultimately the need for metrology-based manufacturing that drives a system towards high precision."

Assembly

Metrology and feedback are the essence of high-precision manufacturing. "The feedback takes what is seemingly impossible and makes it possible and predictable," explained Stack.

In Erickson's career, he has seen optical designs that look as though they cannot be manufactured. "Metrology is the key," he said. "A good test of a high-precision device is that you could not reverse-engineer it and get the same performance. Metrology introduces certain compensations as the system is assembled."

A subtle point here is the difference between making things that are high precision in a laboratory versus a factory setting. "If I was making a one-off device in the lab I may go through many iterations, but this is not cost effective in the real world," explained Stack. "High-precision manufacturing includes the ability to build high-precision assemblies in a factory setting. At a basic level, this is the ability to assemble high-precision devices with skill technicians."

Does my application require high-precision manufacturing?
Although this is a hard question to answer in general, both Stack and Erickson agree that there are some common factors that point towards high-precision manufacturing. The first question you should ask is "does my assembly require active alignment" followed by "is there active compensation involved in the alignment". It is also important to decide if you require the feedback that metrology loops and use of design software, such as CODE V, brings.

Other clues are a design that has multiple axes or multiple paths and has to operate over a wide spectral range (from ultraviolet through to near-infrared). A system running at very low f-numbers that operates at the diffraction limit is also a contender. In other words, there are many factors that point towards high-precision manufacturing but it ultimately depends on the product and its end application.

"In order to make the system work, you cannot rely on the parts alone," concluded Erickson. "You have to look to other techniques that marry the components and allow the whole ensemble to play together."

Example applications
As a metrology expert, Zygo has been involved in a number of high-precision manufacturing programmes and one industry that is particularly reliant on this skill is the medical market. For example, Zygo is a key player in developing high-end medical laser delivery systems for refractive eye surgery.

"The medical laser delivery market requires diffraction-limited beam quality," explained Erickson. "These are often scanning systems that require consistency over a specific depth of focus. Thus there is a volume over which a set level of precision must be maintained."

The first step in the process is to assemble all of the high-precision components that make up the system in what is known as a "dry stack". Zygo then performs interferometry on the stack, feeds that information back into the design code and fabricates compensation spacers that bring the overall lens back into specification – all on the factory floor.

Distortion tester

Zygo has also been involved in the development of production photolithographic lenses. In this example, the diffraction- limited performance and distortion requirements over a large field of view were crucial. "We were working over a large linear field of view requiring 0.5 µm distortion or better," said Erickson. "We built a distortion testing device that can measure distortion in tens of nanometres. We also figured out how to measure the wavefront in-situ and in tandem with the distortion. These lens systems can have between 20 and 30 elements, and will end up in our production facility with volumes of hundreds per year – its no small feat."

One final example is Zygo's work on microscope objectives with a bandwidth of 190–700 nm. "The field of view was relatively small but the spot needed to be diffraction limited," said Erickson. "We also had to control certain aberration groups (certain Zernikes) to achieve the required performance."

As is evident from the photolithographic lens example, Zygo typically builds specialized tools that are customized for the end application. It is also crucial to consider the final environment that the system will be used in.

"Not only do you have high-performance glass, but you also work with high-performance metal in the packaging," emphasized Erickson. "This goes right through from the metal to the adhesive that you are using. Knowing what is going to happen in the field is all part of the high-precision manufacturing process. You may spend as much time working on that as you did on the original lens design."

Cost and other misconceptions
One of the most common misconceptions is that high-precision manufacturing is cost-prohibitive. "There is an interesting distinction between cost and value," said Stack. "It's the value of the part that counts. We have found that the value of higher performance balanced with good manufacturing techniques actually turns out to be a more defendable space."

Other factors that influence costs are tooling and the order in which the system is assembled. "When we develop a prototype, we throw a lot of tools at it," said Erickson. "When we make it in production, we refine the approach using our design for manufacturing and assembly (DFMA) process. We take various things out through the DFMA process and the final order in which you put something together drives cost. We take all of this into account and assemble the system in the right order to make it cost-competitive."

It is important however to build enough time into any product roadmap to allow the DFMA process to run its full course and reap the maximum benefit. "Another mistake that I have observed, particularly in time-sensitive markets, is a lack of appreciation of the time it takes to turn a prototype into a product," said Stack. "It is hard to apply a rule of thumb here as it all depends on the production quantities, but this error ultimately slows down the launch of the product."

According to Stack, after cost, one of the biggest mistakes is thinking that there is a high percentage of completion once the optical design is on paper. "This is a very poor assumption," he said. "The customer's functional specifications and requirements do not necessarily match directly, or easily, to what are typical optical tolerances such as MTF. The trick is to help customers achieve functionally robust products without over specifying the optical requirements. This confidence is almost always achieved by providing highly refined metrology and production techniques at a system level."

This mismatch between the process capabilities and the system requirements is a common problem says Erickson. "Customers often miss the limitations of optical shops and are surprised by the experience as they go through the first system build," he said. "There has to be a high degree of collaboration. As we work through the specifications, there has to be give and take to support the customer's cost requirements, manufacturing capabilities and time to market."

Conclusion
A lot of work goes on behind the scenes in high-precision manufacturing that is not immediately obvious to scientists who have never been through the process. Once you have made the leap of faith, committed to putting your product through this manufacturing approach and overcome the stigma of cost, the rewards are clear. "We are able to streamline things but you will never eliminate the need for the metrology feedback in high-precision manufacturing," concluded Stack. "We keep working the DFMA process and the encouraging thing is that it starts to converge to precision pricing."

• This article originally appeared in the March 2008 issue of Optics & Laser Europe magazine.