Incandescent lamps have almost vanished from many new avionics designs. Now high-efficiency light emitting diodes (LEDs), especially in white, have become inexpensive and readily available.
LEDs are widely used for backlighting pushbuttons, panel overlays, and liquid crystal displays (CRTs), as well as for status annunciation and general indication. Getting the best results requires a good understanding of parts specifications and the important optical properties of these low-cost light sources.
LEDs are not like lamps, and that is both a good and a bad thing. First, they have no surge current at turn on. That is a very good thing and one of their big advantages over lamps, which often have a five-fold increase in load current when first turned on. Second, LEDs come in specific narrow bandwidth colors, which is great if you want color but a nuisance if you really want wide spectrum light (white) for general illumination.
LEDs also have what is widely thought of (incorrectly) as "unlimited life." Their lifespan really translates into half intensity at about 100,000 hours of operation. That’s still a respectable period and better than most lamps, which struggle to reach 25,000 hours before failure.
Finally, LEDs are narrow-angle emitters, and this is often a problem, especially when they are replacement lamps. They just don’t have the same illumination physics, and that can be a lot more troublesome than it first appears.
The burning filament in an incandescent lamp is an almost perfect true spherical emitter, a feature many thought was detracting and useless until technicians tried to replace it with a narrow-beam LED. Lamps radiate uniformly in every direction, making them effective for panel illumination and creating a source of easily reflected light at 90� off axis, a critical parameter for good diffused illumination. LEDs, however, emit in a forward cone, often quite narrow, usually specified in degrees, such a 7�, 15� and so on. They achieve high intensity by reducing the angle through which they emit, and they have virtually no 90� off-axis light.
You can think of LEDs as being represented by X number of photons streaming from a point. If all can go only through a narrow cone, they appear bright and concentrated to the observer. But if they must radiate in all directions, the overall appearance is dim. As the angle of the cone increases, the perceived intensity drops, but the actual emitted energy remains the same.
Many engineers do not realize that the ultra-bright LED may be the exact same chip as a far less powerful one, but designed in a cavity and lens package to give a narrow angle of illumination. Designers often reach for the super bright parts to get better illumination, not realizing they may get the exact opposite result, thanks to the lens and reflector nature of the LED package.
Chip or surface-mount device (SMD) packages tend to give the widest angles of illumination, and forward angles up to 170� exist. But be careful, these may also have tightly focused lenses. Keep in mind that this is still zero illumination at 90� off-axis. This is why backlighting panels takes so many LEDs, one almost literally under every item to be illuminated.
A surprise to designers is that sometimes the LED radiation pattern actually has a dark spot in the center when placed too close to the object being illuminated. This would be humorous if it didn’t have such disastrous consequences for the designer. Likewise, many LED packages have a focal point that is some distance away from the surface, and have very peculiar "close-in" characteristics, which can only be seen in a simple experiment with a piece of paper held up to show LED’s emission pattern.
Speaking from an unhappy experience, I can assure you this test can save you some real grief. It’s true, there really are dark emitting diodes.
Status annunciators made from plain LEDs may also create an off-axis visibility problem. Even brilliant emitters at 2,000 millicandellas (mcds) are invisible when seen off-axis farther than their design angle. And this may be as narrow as 3.5� off-axis. Using LEDs to flood a flat legend from far away (which then radiates uniformly) gives better visibility off-axis, as the diffuser spreads light over a much broader field of view.
LEDs also exhibit a fairly straight-line change in intensity versus temperature. They are bright when cold, dim when hot. Sadly, it’s just physics. You can’t really do anything about it, except hope for cold weather. Over normal operating ranges of 40� to 85�C, a diode commonly doubles and halves its intensity relative to the 20�C test value. This can be a real problem and is a critical parameter for people using LED opto-couplers to remember, if they require high current transfer ratio (CTR) values.
Recently, many people have started using "white" LEDs for general illumination, and they can and do gain amazingly good results. White emitters are really blue (short wavelength) emitters, with a phosphor target in front of the LED die. The target is excited by the high-energy, short-wavelength LED emission, and then exhibits secondary emission over a broad range, giving us the appearance of white light.
The white emitters generally have a blue tint, which often is desirable in the cockpit, but they are missing some wavelengths and will not work in every application, where true color-correct reflection is required. For example, a small, map illuminator that may not show warm colors could make the map reading difficult.
White LEDs also have a less-tolerant low-temperature rating, and many are not specified below 20�C, due to package and materials limitations. You will need to match the devices’ temperature ratings against your target DO-160 survival range. You may be dismayed; many devices will not meet the required temperatures, especially the mandatory 55� and 85�C survival ratings.
Various other quirks emerge when it comes to applying LEDs in your design. Dimming is not difficult but generally does not track the lamps’ behavior well. Cartridge LED lamps in switches and annunciators may also catch designers unaware by actually being series LED clusters, sometimes with very high turn-on voltages (>10V). These can be difficult to integrate into a common dimming scheme with other light sources.
Pulse width modulation (PWM) dimming should be used with care. The modulation can create motion flicker in the LEDs (when someone turns his or her head) if the clock rate is too slow. It also can be a serious source of electrical interference if high current square waves are employed at high frequencies.
One last issue to think about: LEDs are near coherent emitters. They like to produce parallel beams of light. This is the opposite behavior of lamps that radiate in every direction. And many optical issues we take for granted, like diffusion, reflection and penetration, are altered dramatically as a result.
I spoke recently to a manufacturer of lighted overlays, and we were comparing notes on the problems of getting LEDS to produce a nice uniform illumination of a legend, just like a lamp. I believe his exact word was "impossible" unless you get far away. Something to think about in your next lighting design.
Walter Shawlee 2 may be reached by e-mail at [email protected].