From plug-and-play to high performance (continued)

-- Test & Measurement World, 8/1/2006

A continuation of our interview with Brian Doody, COO, Dalsa, which appeared in the August 2006 Viewpoint column.

Click here to read the first part of the interview.

Brian Doody COO Dalsa Waterloo, ON, Canada Brian Doody became COO of Dalsa in 2006, with responsibility for overseeing all of the company’s internal operations, including both the Digital Imaging and Semiconductor businesses. He joined the company in 1985 and has since served in multiple roles, including VP of manufacturing and engineering and VP of operations, as well as president of Dalsa Digital Imaging, where he was responsible for the firm’s $100 million digital imaging business. A professional engineer, Doody has a BSEE from Queen’s University and an MS from the University of Waterloo. In an exclusive interview with T&MW, Doody discussed the challenges facing manufacturers of vision systems.

A. About one third of our revenues in our imaging business comes from Asia Pacific, and we recently established a new business unit to serve that market. We see greater opportunities for new business there than in North America or Europe. In China, where we have just opened a new operation in Beijing, many new factories are employing vision from the outset as a way of ensuring the quality and efficiency needed to compete in world markets.

Q. To what extent are customers looking for plug-and-play vision solutions, such as those offered by your ipd unit?

A. The ipd unit was quite attractive to us and was a major consideration as we evaluated the acquisition of Coreco, which we completed in 2005. Ipd traditionally targeted customers who had not used machine vision and wanted plug-and-play solutions, such as the smart cameras that have become more and more popular over the last five years. This was a major change for Dalsa, which had focused on custom and high-performance applications, but we see big potential in serving this broader market.

In recent years, Coreco had been investing steadily in the ipd unit. They essentially took the software and the algorithm skills from their acquisition of Imaging Technology and integrated them into new camera products. They began to offer compact solutions, typically using one to four off-the-shelf cameras. Yet, this compact package can provide real-time image processing, as well as control hardware on the production floor.

The target market is the factory floor technician who has not used machine vision before, needs simple machine-vision inspection, and does not want to spend a quarter of a million dollars on a custom system. This could be as simple as making sure that all the buttons are in the right places on a keyboard. With the ipd products, you can get a pretty powerful inspection system running in a very short time for $10,000 or less. The result is improved quality of products and reduced labor time.

Q. What is the potential for your XRI-1200 digital imaging processing board, which won a "Best in Test" award from T&MW?

A. We designed this frame grabber to meet the specific needs of x-ray imaging. This FPGA-based board delivers a level of performance that goes well beyond what a general-purpose frame grabber could do. These functions include image capture, shading correction, gamma correction, lens correction, motion detection, and noise reduction. A major application is portable x-ray systems for the medical market, such as the C-arm fluoroscopy systems used in emergency rooms. But there are also some good industrial applications, such as x-ray systems for nondestructive testing and for inspection of printed-circuit boards.

Q. How do you keep engineers up to date on your vision products?

A. In the case of our ipd customers, a major focus is training and education of factory floor engineers and technicians in plants that make everything from detergents to automobiles. A lot of our training for this group consists of onsite visits to plants from distributors of our equipment or from our own application teams. This is very hands-on, and we usually can get systems up and running in short order. For higher-end applications, we post important white papers on our Website and conduct Webinars. We also do training workshops at trade shows and other locations for customers and distributors on new software and other products.

Q. In terms of Dalsa's overall revenues, how do digital imaging products stack up versus your semiconductor lines?

A. In 2005, imaging accounted for about 70% of our total revenues of $167 million Canadian. The year before, semiconductors accounted for about 40% of sales; so it does vary from year to year. In the spring of this year, we announced a reorganization that puts our image sensor component business, which had been part of the imaging group, into our semiconductor group. This CCD chip sells to a different type of customer base than do complete camera units and image processors. As a result of this shift, imaging this year is likely to account for 50-60% of our sales.

Q. MEMS wafer processing is a big part of your semiconductor line. Correct?

A. Yes it is. We acquired the semiconductor business in 2002 from Zarlink, which had a strong intellectual property base in MEMs technology but had not yet brought a product to market. We bought the Zarlink fab to ensure a ready supply of CCD image sensor wafers for our internal use, but we also saw MEMS as being a significant growth opportunity that would open up a new market for us and keep the fab in full production. So, we invested in that technology to the point where our MEMS business is now almost equal to our CMOS business.

The biggest applications for our MEMS products are stand-alone MEMS chips, such as those used for pressure sensing in automotive applications. But we also offer wafer-processing services for our customer's more advanced MEMS chips, such as those that replace microphones and speakers in cell phones. MEMS made at our fab also control fluid flow in inkjet printer heads. In the future, we see good opportunities in the bio MEMS field, where these devices will be used in such applications as sorting cells in a blood samples.

Q. How are you shortening the time from R&D to commercial products?

A. As we grew, we went from developing one or two new products each year to having a large R&D budget—typically 15 to 20% of our sales--that funded a large group of people in several locations who were developing new cameras, new software, and new image processors. About five years ago, we found that our development tasks were too distributed, and there wasn't the same degree of personal ownership of development programs that we had when we were a smaller company. So, we revamped the development process.

Now, we pick a small number of top priority projects every year and staff them with people who have been taken out of their functional unit, such as electronics, software, and mechanical design. The individuals who work in these tiger teams know that their only role is to develop that product. The result is a much greater focus and a shortening of development times from about 24 months to eight or 12 months. In short, we are doing a smaller number of projects a lot better and a lot faster.

Q. How does Dalsa distinguish itself from others in the vision field?

A. Our whole 26-year history and technology development have been based on imaging and vision. Our acquisitions also have related to expanding our capabilities in this field. Right now, we can offer as broad a range of vision solutions as can be found anywhere in the industry. Especially with our newer, higher-volume-market products, such as the Genie camera, we can address virtually any vision application that our customers can dream up. And because we do our own internal development, we can still position ourselves as a high-performance leader for very challenging applications.

Click here to read the first part of the interview.