Uniformity is probably the most misunderstood lighting topic, with many differing points of view.
Certainly, uniformity is important for robust, accurate, repeatable vision applications. No one wants false positives or missed defects.
Fortunately, hardware and software continue to evolve, making uniformity less of an issue, but nonetheless, still important enough. To determine how uniform your lighting needs to be, first review your software's capability, then contact a lighting supplier who will work with you to meet your requirements. Taking this approach will ensure you only buy what you need, and not what someone thinks.
What is uniformity? The topic deserves to be discussed from three perspectives: the fiber product, the light source, and the receiver. All components are required to validate uniformity in a fiberoptic lighting system.
Cameras and associated lenses are the typical receiver set in an actual application. Camera CCD's have sensitivity favoring wavelengths in the 900 nanometer range. Therefore, be sure the method used to validate uniformity employs the use of a sensor with the same response as your system.
The distance from the subject, and the quality of the lens may also affect uniformity measurement. As resolution decreases, (getting farther from the subject) the ability to resolve uniformity decreases.
Focal length, lens flare and spherical aberration may also influence uniformity, but to a lesser degree than the lamp and the fiber part.
THE LAMPNo two lamps are created equal. While some manufacturers tout lamps made exclusively for fiberoptic lighting, the vast majority of fiber suppliers use lamps made for slide projectors. These lamps were never made to focus on a fiber bundle.
Lamps have "hot spots". Some areas within the lamp produce more light than others. The filament's shape, its condition (age) and the lamp environment all contribute to affect uniformity, which is an ever changing characteristic throughout a lamp's life.
Lamps have different N.A. and focal distances. Using lamps with small concentrated spots at the focal point are great for intensity and uniformity, if the fiber bundle is sized for the spot. Most times, they are not, which exacerbates the problem at the output.
Fiberoptic manufacturers use techniques to minimize the impact of non-uniformity in lamps, and they're fairly effective within certain guidelines:
- Avoid using a fiber bundle with active diameter greater than the lamp's spot; doing so affects uniformity at the fiber output. Use multiple light sources, change to a different lamp, or defocus the lamp. (If you can stand the change in intensity)
- Insure the lamp spot is optimally focused on the input.
- Randomize the placement of the fiber at the input to balance the lamp's hot and cold spots at the output.
THE FIBEROPTIC COMPONENT
Skew and the finish quality can affect uniformity at the input and/or the output. Method of construction, care taken during construction (broken fibers) and the design of the part usually influence the output end of the part.
The best design will not overly bend the fiber. It will have a perfectly randomized input with no skewed or broken fiber, and the output fibers will be perfectly straight, nested, and polished. Now the bad news. This is not possible!
If you're convinced the fiber in your application is causing problems, use an integrating sphere to check the output. You'll probably find the fiber alone is not the sole source of your problem.
To summarize: The fiber part, light source, and receiver all affect uniformity. Work with your supplier to develop a specification for your exact application. Don't waste your time comparing different manufacturer's specs. They weren't developed the same way. Provide or purchase a test fixture set up to duplicate your application for the best results.
Overall, the most important consideration is how the lighting system works in your application, not published specifications. However, if you are currently relying on a published specification, you should find out how it was developed and if its claims can be transferred to your specific set of conditions.
For example, a stated specification may be developed for a transmitted application, with a detector held in contact with the output. If your application will be using the lighting product in contact with the photodetector, then the specification has value and meaning. If your application will use reflected light from a subject, imaged by a camera lens at some distance and angle, then the specification probably doesn't have value at all for you, not even to ascertain consistency part to part.
GRAY SCALES vs. PERCENTAGE
It can be confusing,but it doesn't need to be. Gray scales and percentages are two different ways to state the same result.
If a specification is stated in gray scales (256 distinct intensity steps), find out what mean intensity value was used, then divide the gray scale value into the mean value. The result is converted to a percentage. For example: "+/- 10 gray scales at a mean of 200" is the same as stating uniformity of +/- 5% (200 divided by 10 gray scales). Furthermore, +/-5% is the same as stating uniformity within 10%.