The Contact Injector allows a user to inject light into one or more optical fibers. We have two A2080 models with 16 and with 36 LEDs. It is best to pair a contact injector with a plate that holds fibers in place at a fixed distance from the LEDs. If used incorrectly, a fiber will not capture the maximum possible amount of light from an LED.
Different LEDs of the same type output different amounts of power into a fiber. Differences from one LED to another exist not only in maximum total power output, but also in which areas of the LED put out the most light. Each LED has a unique profile of variations in power output along its surface, such as the ones shown below.
This shows a heat-map style chart to show the bright and dark sections of many seperate LEDs. Some patterns are common, such as dark stretches through their centers, but each LED shows a distinct and unique distribution of power output. The differences in intensity output between the brighter and darker patches in similar structures vary as well. The first image above shows the intensity profile of an LED with a dark stripe directly through the middle, as well as large variations in intensity along the top left corner. The second image shows little variation from one point to the next throughout the center. In the first image, the dark stripe through the middle only emits about 35% of the maximum power output at any point on that LED. Drastic differences between areas such as this above were rare, but must be accounted for.
These measurements were taken using a photodiode to measure the light transmitted through an optical fiber. The injector end of the fiber was pressed directly against the surface of the LED The optical fiber could be moved in the x, y, and z directions, where the z direction is perpindicular to the surface of the LED and the x and y directions are parallel to its edges. In the above graphs, the z distance was kept at 0. The intensity profile of any given LED is smoothed out considerably by backing the injector end of the fiber away from the surface of the LED.
When the fiber is not pressed against the surface of the LED but still within a certain maximum distance (derived in the next section), the brightest spots of the LED above are still off to the side, but the relative variation from one place to the next around the center is drastically decreased. The fiber can now "see" parts of the LED that are both relatively bright and relatively dark, so the total power captured is neither the maximum possible nor the minimum possible. Using fibers a small distance from the LED allows the user to avoid concerns about large changes in intensity due to properties of the LED itself. This also allows some leeway for slight imperfections in alignment.
Up to a certain distance, pulling a fiber away from an LED makes little difference in the amount of light captured in a fiber. This is because though the intensity of light reaching the ferrule decreases with the inverse square of the distance, the circular area of the surface of the LED which is visible to the fiber increases with the square of the distance.
Taking advantage of this actually allows the user to reduce the effects of different levels of power output on different parts of the LED, as discussed before. Accepting light from a larger area increases the likelihood of capturing the areas of highest intensity. This is particularly true of LEDs that show the dark central line pattern that we commonly found in our tests. In these situations, if the ferrule is perfectly centered on the LED and pressed up against its surface, the light captured could be far less than the maximum output.
The distance to which this is true depends on the numerical aperture of the fiber used. Our fibers have a numerical aperture of .22, meaning that light can enter the fiber from angles up to 12.7°. The Deep Red Luxeon Z's which we tested are 1mm square, so the maximum radius of the circle visible to the fiber containing only LED is .5mm. The farthest distance between the fiber and the LED that will maintain maximum intensity captured is therefore z = .5mm/tan(12.7°) = 2.21 mm. A more detailed description of the calculation and testing of this can be found here. To account for this, we have produced injector plates for both 16- and 36-LED models which hold fiber tips 1 mm from the LEDs mounted on a contact injector.
When used for too long, an LED will heat up and this will cause it to output less intensity as time goes on.
For repeated short flashes, there is no noticable change in power output due to temperature increase. We tested an LED flashing on a loop for 10ms on, 100 ms off and 100ms on, 100 ms off using a photodiode connected to an oscilloscope and saw no change in intensity output either during a single flash or between multiple flashes. For prolonged usage, an LED heats up and causes the power output to decrease.
The initial output when the LED is turned on continuously is the same as the power output during a short flash, even when the flashes are repeated.
After about 25 seconds, the intensity output drops to 90%. Power output never drops below 75-80% of the maximum because the temperature increase slows down.
Even clean and scratch-free, there will be differences between the intensities of light transmitted by individual fibers. The greatest difference we found between maximum intensities for the same LED was 44.5 uW and 33.8 uW, a 24% difference.
In the above image, fibers 2 and 3 had been protected with plastic ferrule caps whereas fibers 1 and 4 had not. The unprotected fibers were cleaned with a microfiber fabric before use. Fiber 1 was also tested before being cleaned, and little difference was found.
It was clear after each test that the apparatus built to hold the ferrules in place left a small amount of residue on the surface of the ferrule. Heavy scratching will affect the power captured, but if a ferrule is handled properly and used gently, the power capture will not be greatly affected. The use of ferrule caps is still highly recommended, particularly during transport, to prevent scratching.
We found no correlation between the voltage and current applied to an LED and its intensity profile.
Shown above, the same LED is mapped at three different levels of current, and the same patterns can be seen. In all of them, there is a dark trench through the middle and bright patches towards the corners. These images show only cross-sections of the LED through the middles, as a well-centered fiber should not be affected by the intensity differences at the corners. These maps were created by measuring the light output by a fiber whose end was pressed against the surface of a Deep Red Luxeon Z.