I'm trying to couple a white LED to a fiber bundle that's comprised of several 30 uM silica fibers. The bundle outer diameter is on the order of .5 mm. The NA of the fiber is .87 and the whole bundle is placed in a stainless steel ferrule that's .125" in diameter. I've read about different methods of coupling -- everything from a ball lens to gluing the fiber directly over the LED die.

I'm curious if anyone has experience doing something like this. Can anyone delineate the advantages/disadvantages of the various methods? Are the main tradeoffs cost vs transmission efficiency? If so what methods are on which end of the spectrum?

It's worth stating that I'm not sold on any particular LED, but I have experimented with simply holding the ferrule over a LXML-PWC2 -- tons of light comes in (enough for my application) with iF ~400mA. This would be a reasonable solution except that the efficiency of transfer is terrible (obviously) so I have to use more current than I'd like. The result is a rather large heat sink, which I'd like to shrink or ditch all together.

  • \$\begingroup\$ Could you specify what the application is, or at least how much light is required, and what spread is needed at the viewing end? Model railroaders often use end-emitting fiber tightly glued to the head of a lensed discrete 3 mm LED, or even a small-angle SMD LED, using a transparent glue and tight bonding (no air at the junction). At the output end, a soldering iron pressed into the tip forms a little mushroom which emits light well. \$\endgroup\$ – Anindo Ghosh Aug 28 '13 at 11:24
  • \$\begingroup\$ How many fibers in the bundle? \$\endgroup\$ – The Photon Aug 28 '13 at 14:38
  • \$\begingroup\$ The application is to illuminate a small volume -- 1.5cm^3. I don't really know how to specify the amount of light that I need other than qualitatively to say "this looks good." Frustrating I know... With the high power led I get roughly the right amount of light out. If I had to guess as to efficiency of transfer I'd ball park it at less than 10%. I suppose from there I could back into an answer and say that @ iF = 400 mA the led is doing roughly 150 lumens. If I assume 10% transfer then I could call it ~15 lumens of requisite light. That's a pretty major swag though... \$\endgroup\$ – Doov Aug 28 '13 at 18:43
  • \$\begingroup\$ @ThePhoton I believe there are roughly 50 fibers in the bundle, but I will double check. \$\endgroup\$ – Doov Aug 28 '13 at 18:44
  • \$\begingroup\$ Few clarifications -- there are indeed ~50 fibers, but the NA is higher than I thought -- .87. Additionally the diameter of each fiber is smaller than I thought -- 30 uM. \$\endgroup\$ – Doov Aug 28 '13 at 19:00

"Typical" fiber for communication is 9, 50, or 62.5 um diameter in the core, but there is a 125 um cladding that is also necessary for the fiber to work. There is also 900 um core plastic optical fiber. What is out there for illumination, I'm not sure. Clearly you're not working with one of the types that I'm familiar with.

But, the key point is that the core is smaller than the cladding. And the light coupled into the fiber will be (best-case) what falls on the core. With .87 NA you'll probably get pretty close to this ideal. So I'll just make up some numbers. Say you have a 30 um core in a 60 um cladding. Overall the maximum coupling efficiency you could expect from uniformly illuminating the end of the fiber bundle is 302/602 or 25%.

Then there's a factor for the packing density of round objects into a an area (the gaps between the fibers in the bundle) which I believe is about 78% best-case.

And a reflection loss of about 4% for light entering the glass from air.

Add all these up, (.25) x (.78) x (.96), and you have about 19% best-case coupling efficiency. (You'll need to re-calculate this knowing the correct factor for the first term)

If you are getting near this, I'd say you are actually doing pretty well.

Of course it wouldn't hurt to find an LED that emits in a narrowish cone instead of over half of all space, or even to add some kind of lens to focus the light on the area of the fiber bundle. But generally your best case coupling is still not going to be all that great.

| improve this answer | |
  • \$\begingroup\$ Thanks for the answer. Is there a good way to measure coupling efficiency? When I ball parked less than 10% it was simply based on my (inaccurate) swag and how much light was coming out the fiber as to how much light was coming out the base/connection. It's quite possible I'm down at something like 3% in which case 19% would be quite an improvement -- at least I'd be able to lower iF on the diode and loose some heat. What kind of lens solutions exist and what are the tradeoffs? Thanks! \$\endgroup\$ – Doov Aug 29 '13 at 17:26
  • \$\begingroup\$ Measuring the output of the fiber bundle should be pretty easy if you have the right gear. Simply bring the bundle as close as possible to a large-area photodetector. These are readily available with diameters well above .125". Measuring the output of the LED itself is harder --- if you get a big enough photodetector and hold it close enough to the LED, you'll probably get a reasonable estimate of the total available power. \$\endgroup\$ – The Photon Aug 29 '13 at 19:14
  • \$\begingroup\$ Instrument-grade sensors of this kind (and the meter electronics to read the response, calibrated in uW) are available from companies like Newport, ThorLabs, etc. If you want to do it on a tight budget, I'm not sure exactly where to go. \$\endgroup\$ – The Photon Aug 29 '13 at 19:15
  • \$\begingroup\$ Thanks a lot for the info! Any suggestions for a lens to use just as a comparison to straight mechanical coupling? \$\endgroup\$ – Doov Aug 30 '13 at 18:46

Well your choice of a white LED is the worst possible choice you could make, unless you simply have to have white light for your application. The reason is that white LEDs have significantly lower luminance than colored LEDs, because the light is being emitted from a phosphor that has a large surface area, compared to the bare LED die of a colored LED.

A lens is not going to help much, unless you can get a bare die LED without bond wires to get in the way of the optics, and you will have a hard time finding a small enough lens with a large enough operating NA and low enough aberrations to actually work properly. You would have to focus a real image of the bare die, onto the end of the fiber bundle. But you can never get the luminance of the image to exceed that of the source. (second law prohibition).

Best bet (assuming you really have to have that white light), is to find a small flat surfaced LED (chip like) and glue it onto the end of the fiber bundle with a clear glue, with the smallest possible spacing between the fiber and the LED.

| improve this answer | |
  • \$\begingroup\$ I need white light. Hence the white LED. Why will a lens not help? My understanding is that a ball lens isn't a bad idea for this application. Why would a flat led be better? Just for the sake of adhering it? Generally speaking they have much wider dispersion angles (so I'd assume less light directed towards the fiber). \$\endgroup\$ – Doov Sep 6 '13 at 1:42
  • \$\begingroup\$ Well two things stop light from going into your fiber bundle. One is the surface area of the fiber (core) and the other is the numerical aperture (NA) of the fibers. If U is the maximum angle of incidence on the core, that can be transmitted, and H is the diameter of the core, the quantity NHSin(U) is an invariant under all optical transformations. It is called the Optical Sin theorem, or sometimes the lagrange invariant. The only way to get more light in is to increase the radiance (or luminance) of the source. A lens can't do that a ball lens is bad because of aberrations that lose more \$\endgroup\$ – user28555 Sep 6 '13 at 4:13
  • \$\begingroup\$ Maybe I'm not understanding -- why can't a lens help collimate light into the fiber? I have to imagine that there's a third thing stopping light from entering the bundle -- the direction of light relative to the bundle. After all if I rotate the led 180 degrees from the bundle then zero light goes in no matter surface area/NA. I don't understand why a lens can't help -- If 2 leds emit the same number of photons, but one emits them over a smaller area than the watts/m^2 in the smaller area case is > than the big area case. Can't a lens help focus the light to a smaller spot (e.g. my fiber)? \$\endgroup\$ – Doov Sep 6 '13 at 6:57
  • \$\begingroup\$ Perhaps I'm not understanding this, but checkout the FAQ from edmuds optics "I'm looking for the best way to maximize the amount of light from an LED into a fiber optic. What are the best options for coupling my LED and fiber?" edmundoptics.com/technical-resources-center/… \$\endgroup\$ – Doov Sep 6 '13 at 6:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.