I am part of the small group of people around the world who really want to replace any and all lighting technology with LEDs. Don’t misconstrue it for insanity, but a desire to make aircraft lights more reliable. But then, a nagging question crosses my mind : At what cost?
I have been seeing airplanes flying overhead, in the night sky. And then I have seen LED lights being designed and made at the company I work at. And every single day, every minute that I spend designing these lights, I wonder if they will be just as “effective” as the flashtubes.
Flashtubes, as experienced lighting designers will know, produce a tremendous amount of light. They produce so much light that most aircraft strobe lights need no secondary optics (reflective surfaces or refractive media) to meet the desired intensity distribution pattern as required by the FAA. They throw light everywhere, and they pump out a tremendous amount of light in a very short duration.
How short? The discharge of a flashtube lasts for around 1–10ms. Here is an timing image by Goodrich which shows for how long a flashtube’s flash lasts. Also note that the intensity of the flash varies with time, rising to a peak before decaying to zero.
There must be a method by which the FAA determines the intensity of a strobe light, and then after determining the intensity, deems if the light meets its minimum specifications or not. The method, and the intensity requirements (for FAR Part 25 operators), are outlined below:
Determination of Effective Intensity:
Ie = effective intensity (candles).
I(t) = instantaneous intensity as a function of time.
t2-t1 = flash time interval (seconds).
Normally, the maximum value of effective intensity is obtained when t2 and t1 are chosen so that the effective intensity is equal to the instantaneous intensity at t2 and t1.
Minimum effective intensity requirements:
If the peak intensity of the flashtube discharge is Ipk, and the duration for which the flash is considered “on” is Tf, then the area under the curve may be approximated as a triangle, whose area is:
1/2 * (Ipk * Tf)
This serves as the numerator for the mentioned Blondel Rey equation. If 400ecd (effective candela) minimum is to be met,
400ecd = 0.5 (Ipk * Tf) / (0.2 + Tf)
Since Tf is of the order of 0.001ms, it may be ignored in the denominator.
Thus, to meet a minimum of 400ecd, the flashtube must give out a peak intensity of:
Ipk = 2*(400 * 0.2) / (Tf)
For Tf ~ 1ms,
Ipk = 160,000cd
Goodrich has a product for the Airbus A320 aircraft. The wingtip white anti-collision light, 2LA 005 308-05, produces 2,000 effective candela of light. This intensity is met by a parabolic reflector placed behind a flashtube, the flashtube producing around 800,000cd peak intensity. (for the calculations presented above)
Just for comparison, each of the Airbus A320’s powerful landing lights (PAR 64 using a GE Q4559X consuming 600W of electric power at 28V DC) produces a steady 750,000cd peak intensity.
LED Strobe Lights
Most LED strobe lights in the market today have a flash duration of between 250-300ms. The equation used by all manufacturers today is still the Blondel-Rey equation. According to this equation, the effective intensity of the flash will be:
Ieff = Is*(Tf)/(0.2+Tf)
The integration is gone from the numerator as the flash is of constant intensity. “Is” is the steady state intensity of the LED light.
Ieff = Is*0.3 / 0.5 = 0.6*Is
In other words, one only needs to design the LED light to provide a steady state intensity of
Is = Ieff/0.6 = 1.667*Ieff
For 400ecd, this reduces to
Is >= 666.67 cd
So which is brighter?
If one were to evaluate the intensity-time product of the LED, and compare it with the time-intensity product of a flashtube,
LED = 666.67cd X 0.3s = 200cd-s
Flashtube = 160,000cd X 1ms X 0.5 = 80cd-s
(The 0.5 accounts for the near triangular temporal flash waveform)
It would appear that the “energy” in an LED based light is 150% more. But strangely, according to the Blondel-Rey equation, both produce the same effective intensity.
But as you, pilots, manufacturers and I know, a flashtube is a lot more “catchy” than a longer duration LED flash. What is responsible for this catchiness? A flashtube that produces a lot more light than a 400ecd system requires? Or is it the punch delivered in packing a tremendous intensity in a short duration, akin to a quick pin prick as opposed to a firm press?
The human eye is complex, and even the best of researches admit to knowing little about it. Most experiments conducted, regarding flashing lights, have, in my humble opinion, a small pool of test subjects: the likes of 3-5. Out of these tests come some good revelations, but what does the average human population say? Let’s see some perceptual phenomena in place today, while also diging a little deep into the Blondel-Rey equation that dominates section 1401 of all FAR operators.
1. Broca-Sulzer Effect
A brief, relatively bright flash of light (optimal flash duration of 0.05 to 0.1 s) is subjectively perceived to be brighter than a longer flash of greater luminance intensity.
2. Brücke-Bartley Effect
Below the critical flicker frequency (i.e., the frequency where a flashing light appears constant), the apparent brightness of a flashing light will gradually increase as the frequency is reduced and reach a point (approximately 8 to 10 Hz) where it appears brighter than an uninterrupted light source of equal luminance
3. Bloch’s Law
For sufficiently short stimulus durations, detection threshold decreases inversely with the duration of the stimulus. (Under about 100 milliseconds stimulus duration it is possible to exchange the amount of light for the duration and maintain a constant effect.)
4. Blondel Rey
The product of flash intensity times its duration is equal to the asymptotic threshold value times the sum of the duration plus a “visual response time constant”.
The problem with the Blondel-Rey equation is that, it is based upon experiments conducted at the threshold of visibility. This would mean very dim lights, which were then modulated in pulse to be determined by an observer as visible, or not.
But such is not the case with aircraft anti-collision lights. The lights are visible over long distances, extending to around 10 miles on a dark clear night. But all we need is visibility within 2-3NM, during approach, and take-off. For the other phases of flight, TCAS and ground based radars will be more than sufficient in early warnings, and in high workload environments where sufficient attention is not paid to what’s outside the cockpit.
Quoting D Couzin, ” In the early 1900’s, detection of just visible flashes from lighthouses had the greatest practical importance. Today, conspicuity of very visible flashes has the greatest practical importance.”
Infact, the Blondel-Rey equation conflicts with the findings of the Broca-Sulzer Effect. If a 50ms-100ms duration flash is perceived brighter than a steady state light, then the Blondel-Rey equation which suggests that a 50ms – 100ms light will appear 0.2 – 0.33 times the steady state glow is flawed. I have not conducted experiments on this till date, but when I work with LED lights, I can sense that lights with a shorter flash duration have a greater effect on the eye, though they don’t hurt as much.
If findings, either by way of my own experiments, or the work of men of authority like Y Ohno and D Couzin do reveal that the conspicuity of anti-collision lights has to do with the flash duration, then a serious question will be raised:
Are LED lights too long in duration to serve their purpose?
Watch this video, of an intense, bright flashtube being compared with its replacement LED light. Note that the flashtube throws an awful lot of light, illuminating the left portion of the gentleman’s face, while also his jeans. Note when the Aveo LED flashes, it throws almost nothing. And because the high intense flashes from the flashtube are of a short duration, they catch the eye a lot better than these LED lights ever can.
Another Example is that of the new Boeing 747-8. This aircraft uses LED wingtip White strobes (from Honeywell), and Flashtube based fuselage Red Anticollision lights (again from Honeywell). Note how the Red Anti-collision is far more conspicuous than the White LED lights: