Collimating flash into fiber optic guide

Pau · Post by **Pau** » Wed Oct 05, 2011 9:25 am

For microscope illumination I use a fiber optic guide with a fiber bundle diameter of 6mm placed at few mm of the Canon EOS built in flash. It conducts the light to the microscope illumination optic train.
This setup is inspired in the Graham's one
http://www.photomacrography.net/forum/v ... 2056#22056
Of course I only capture a small fraction of the flash ouput. It is fully adequate for brightfield and also works for darkfield, cross pol and DIC but I want to capture as more light as possible to short the flash duration. At high magnification sometimes I need to use 400 ISO and there is some motion blur due to the shutter, and I plan to put a beamsplitter to be able to have simultaneous incandescent and flash light.

I want a lens system able to capture most of the flash output and concentrate it in a 6mm circle with aproximately parallel rays because the fiber optic aceptance angle.
This lens need to be compact, inexpensive and not difficult to find

I've tried microscope condensers but without much success, surely because its divergent light output
I also had got this lens but don't deliver any benefit:
http://www.photomacrography.net/forum/v ... 2379#22379

Any help will be much apreciated

lauriek · Post by **lauriek** » Wed Oct 05, 2011 11:05 am

I tried to use a rodenstock 50/0.75 to do exactly this but it didn't work very well. I'll be very interested to see what suggestions come up in this thread!

Charles Krebs · Post by **Charles Krebs** » Wed Oct 05, 2011 12:08 pm

Pau,

This is something I played around with off-and-on when I too used a FO light guide for flash. Unfortunately I never came up with a small, simple solution that offered significant benefit over simply placing the end of the light guide as close as possible to the flash tube.

I'm a little surprised that you experienced no benefit with a microscope condenser. I did not try it but it seems like a promising approach

. microscope condensers but without much success, surely because its divergent light output

Probably correct. Fiber optics do have an acceptance angle (NA). Perhaps an addition lens could be used on top to the condenser lens to create a less divergent, more"parallel" beam. I've often thought that a ball lens might figure into a possible solution. See:
http://www.edmundoptics.com/technical-s ... /?&viewall

Looking at that page, think the "reverse" of Figure 3 (i.e. the microscope condenser, with its diverging beam on the right, and the FO guide tip by the dimension labeled "d". (These lenses don't seem to be readily available large than about 10mm diameter, which might be a limiting factor).

Cactusdave · Post by **Cactusdave** » Wed Oct 05, 2011 2:56 pm

A simple 'snoot' of highly reflective material terminating in a hole just large enough to accept the end of the light guide might be quite effective.

Pau · Post by **Pau** » Thu Oct 06, 2011 8:22 am

Many thanks for the comments and sugestions

Laurie, it seems an unsolved problem for some years now

Charles, not very encouraging that you tried it before without succes. A small ball lens at the focus point of the condenser may be an option, but because condensers are designed for almost parallel light rays entrance it remains unclear if it is capable to collect highly divergent rays.

Dave, I did try with a matte silvered cardboard cone (I need to try with a more reflective material) but to have any possible gain it need to be very long in order to produce reflections at very high angles. Even then, I'm not sure fit the light intensity is really higher than close to the light source. This will mechanically interfere with the microscope stage, so it isn't practical for the built in flash but perhaps would be adequate for an external one.

lauriek · Post by **lauriek** » Thu Oct 06, 2011 8:28 am

If you're cone is too long then the inverse square law means too much light would be lost by the gap distance between the flash and the end of the light guide (I would think!)

vendav · Post by **vendav** » Thu Oct 06, 2011 12:05 pm

Laurie,

I'm no expert, but if the cone were to be silvered, would the inverse square law be applicable - where would the light be lost to?

david

Charles Krebs · Post by **Charles Krebs** » Thu Oct 06, 2011 12:39 pm

Charles, not very encouraging that you tried it before without succes.

Don't let that discourage you... I've tried plenty of things without success!

One reason I did not pursue it too hard is that I was satisfied (at the time) with the amount of light I was getting with the simple method. But I used a "bifurcated" ("Y" shaped) light guide in reverse. The flash went into one part of the "Y" the ambient light (from a FO illuminator) went into the other. So I had no light loss from a beam-splitter to deal with.

Some additional "reading". (Look at the last paragraphs of "Application 6"):
http://www.newport.com/Fiber-Optic-Coup ... ntent.aspx

http://www.newport.com/Focusing-and-Col ... ntent.aspx

http://www.newport.com/store/genContent ... 81845/1033

DQE · Post by **DQE** » Thu Oct 06, 2011 7:00 pm

One line of analysis that seems to be applicable to flash snoots:

Each time light bounces off a surface, one is left with X percent of the light. For a 90% reflectance surface, one is left with 0.9 of the incident light per bounce. If light bounces an average of say 5 times on its way out of the snoot, one is left with (0.9) raised to the 5th power, which is about 59 percent of the flash-emitted light.

If the flash is collimated light instead of diffused, the math would seemingly be different in each case, but the inner surface reflectance losses would remain, at least qualitatively.

I'm sure there are computer programs out there that would calculate a fairly accurate total light transmission value for snoots.

I seem to recall that more sophisticated light transmission devices exist, featuring things like total internal reflection technologies, fiber optic light guides, etc, etc. They would of course work very differently and may be much more efficient.

I hope these ad-hoc thoughts are useful or at least interesting to some degree! Efficient snoot and flash diffuser design is certainly of interest to most of us.

AndrewC · Post by **AndrewC** » Fri Oct 07, 2011 11:58 am

DQE wrote:One line of analysis that seems to be applicable to flash snoots:

Each time light bounces off a surface, one is left with X percent of the light. For a 90% reflectance surface, one is left with 0.9 of the incident light per bounce. If light bounces an average of say 5 times on its way out of the snoot, one is left with (0.9) raised to the 5th power, which is about 59 percent of the flash-emitted light. ..

Yes, but that is 59% of light that would otherwise be completely lost !

soldevilla · Post by **soldevilla** » Sat Oct 08, 2011 3:21 am

Charles Krebs wrote:But I used a "bifurcated" ("Y" shaped) light guide in reverse. The flash went into one part of the "Y" the ambient light (from a FO illuminator) went into the other. So I had no light loss from a beam-splitter to deal with.

ChrisR · Post by **ChrisR** » Sat Oct 08, 2011 3:30 am

Well to be fair it's sort of half

, with my light guide. It didn't work quite as well as I hoped when I tried it, back along.
One little-end's fibres are all on one side of the big end's fibres so they aren't quite as coincident as would be nice.
Sort of

really !

Pau · Post by **Pau** » Sat Oct 08, 2011 8:45 am

Laurie, David: If the cone really worked as we want the square law woldn't apply, but in my very unaccurate and limited tests both effects (beneffit of colecting light and loss due to distance) seems to cancel.

Phil, interesting points. Total internal reflection is about VERY small incidence angles. This, and the need to minimise the amount of reflections, is why I think the cone needs to be long, but I don't have the skills nor the software to do the calculations.

Soldevilla, the Y bifurcated fiberoptics setup is illustrated in the Graham's post I linked and in his website. I have one of those but it isn't useful because I have the same problem that Chris explained. I'm aware that some makers do this kind of guides with ramdomized fibers but most are half and half in the common part. If I find one of these with enough diameter it would be better than the beamsplitter approach I'm going to implement.

Charles, thanks for the links. The literature seems mainly oriented towards laser aplications. Was you Y guide randomized?.
I'm in a point trying to decide as you did between following with the fiber optic approach that seems more uncertain and to copy yours, better experimented

but more expensive and difficult for me to implement. (I don't have the parts and I would need the help of an electronics skilled friend

ChrisR · Post by **ChrisR** » Sat Oct 08, 2011 9:09 am

"Charlie's" articles are interesting, but seem to miss out the one I expected would be useful for our application. (unless I'm being old) It would be the reverse of this one, which is intended for expanding a laser to a wider beam. We want to collect from say a parabolic reflector which catches all the flash output, and squish it up a fibre. So imagine the flash on the right.

http://assets.newport.com/web250w-EN/images/1272227.gif

In practice, flash power is equal to a very long time exposure from any "bulb" that I've tried, so I'm thinking we don't need particularly good coupling between the flash and the fibre.

By the way, T6 LEDs, which aren't wildly expensive, are much brighter than those Ikea Jansjo lamps we use. Conversion kit, anyone??

ChrisR · Post by **ChrisR** » Sat Oct 08, 2011 9:26 am

Pau, before you arrived there was some chat about light guides, acrylic and the like.
It was easy to Search out, because I could use the word Ballcock.
I wonder if Mrs Ballcock ever finished her bottle of Turkish Bath Balm. (cf page 3)
Here!