Has anyone read any new LED in research?

Could you just carefully sand and or polish the led dome flat or is it not thick enough?

While sanding it flat you are very likely to knock the dome off.

DM51 said:

Maybe you mean this thread by Ra?

Click to expand...

That's the one, DM! Good memory. That saved me a little time as well since I was going to do a search for it.

It would require someone such as Ra actually being able to produce the desired product as described by 2xTrinity (this one sounds a bit complicated), then convincing him it'd be worth his while to do a run. A custom TIR optic for the purposes mentioned here would NOT be cheap! Then again, there are a number of people here who are able, and possibly willing to pay for the new world's longest throwing single LED torch (so, the Databank70 doesn't count... lol).

Yes, this brings the thread even further OT, but man, I'd love to see an optic like that born of the ideas and abilities of those here on CPF.

Bimmerboy said:

Actually, one guy here machined his own high quality glass optic which was quite a slick little piece of DIY engineering. I don't remember who it was, but he may have worked in some optical field (ha... get it?... optical field. A little opthamological humor there) for a living, and/or had access to the proper equipment.

If the thread can be dug up, it's possible there's a way to contact him. If there's enough money in it, there's usually always someone willing to do a run of whatever custom item.

Click to expand...

Yeah I remember that one. Just because he can do a normal optic does not mean he can or wants to do this new optic. I may be wrong but this seems to be a much more difficult task than normal TIR optics. Those optics like the one he made have a hole that goes up the middle and is used to redirect light so that it can be collimated properly. If you fill that in with optical gel or similar it will not work right. So basically it is a no-go on that design as far as I'm concerned.

Actually machining cast acrylic into an optic is not all that difficult with a CNC mill and/or lathe.

Oznog said:

Actually machining cast acrylic into an optic is not all that difficult with a CNC mill and/or lathe.

Click to expand...

Yes but figuring out exactly what shape it should be is.

I don't think that would be too hard with a bit of experience, math, and some ray tracing software..

This thread belongs in LED rather than LED Flashlights, so I'm mving it now.

Click to expand...

Considering all I'm seeing is a discussion about flashlight optics, why did you bother? rant/off

For a second I though the topic was about improved LED technology for practical uses - fooled again. I'm also guessing the new LED technology talked about here is just more obnoxious, cold white emitters with a CRI lower than my golf score.

Ok I was doing a little more research into this and again did the same experiment only this time using oil. Why oil? Because the refraction index of corn oil is 1.47 which is extremely close to the refraction index of the glass (1.50). Immersing the LED in oil causes the surface area to shrink by an aditional 10%. In all what this means is the apparent surface area of the die is 34% as large as when compared to the non-immersed LED. Basically What I'm saying is that by immersing the LED in oil the light becomes 3x more concentrated.




By photographing the LED submerged in corn oil, it suddenly becomes clear the shape of the cree ring around die. The oil completely undid the light distortion effect caused by the glass dome.

I'm wondering what would happen if an LED were immersed in a substance with a refraction index greater than glass. Do you think it would cause the light emmited to become even more focussed than the 1mm die?

blasterman said:

Considering all I'm seeing is a discussion about flashlight optics, why did you bother? rant/off

For a second I though the topic was about improved LED technology for practical uses - fooled again. I'm also guessing the new LED technology talked about here is just more obnoxious, cold white emitters with a CRI lower than my golf score.

Click to expand...

Aside from gratuitous rudeness, what exactly was the purpose of that intervention? If you can't remain civil, stay out of the discussion. If I see another post from you similar in tone to that one, your activities will be suspended.

jashhash said:

I'm wondering what would happen if an LED were immersed in a substance with a refraction index greater than glass. Do you think it would cause the light emmited to become even more focussed than the 1mm die?

Click to expand...

Unfortunately, can't get more focused than the 1mm die at least not effectively as it would violate Etendue http://en.wikipedia.org/wiki/Etendue

Semiman

jashhash said:

Ok I was doing a little more research into this and again did the same experiment only this time using oil. Why oil? Because the refraction index of corn oil is 1.47 which is extremely close to the refraction index of the glass (1.50). Immersing the LED in oil causes the surface area to shrink by an aditional 10%. In all what this means is the apparent surface area of the die is 34% as large as when compared to the non-immersed LED. Basically What I'm saying is that by immersing the LED in oil the light becomes 3x more concentrated.

By photographing the LED submerged in corn oil, it suddenly becomes clear the shape of the cree ring around die. The oil completely undid the light distortion effect caused by the glass dome.

I'm wondering what would happen if an LED were immersed in a substance with a refraction index greater than glass. Do you think it would cause the light emmited to become even more focussed than the 1mm die?

Click to expand...

The effective die size increases because of the way the light is refracted from the dome. Refraction takes place whenever there is a transition between media of different indices of refraction.

Whether that is from a high index medium down to a low index medium, or the other way around (eg, glass immersed in some higher index medium) -- the amount of refraction (and also, unwanted reflections from these interfaces) will still be more the greater the index mismatch, in either direction.

The corn oil/glass match there is as good as you can possibly get.

Yeah I remember that one. Just because he can do a normal optic does not mean he can or wants to do this new optic. I may be wrong but this seems to be a much more difficult task than normal TIR optics. Those optics like the one he made have a hole that goes up the middle and is used to redirect light so that it can be collimated properly. If you fill that in with optical gel or similar it will not work right. So basically it is a no-go on that design as far as I'm concerned.

Click to expand...

In order for the index-matched type of TIR optic to work right, it will necessarily need to be deeper, and have a convex outer surface. This is because the way that RA designed his optic, there is actaully an aspheric lens INSIDE the cylinder where the LED resides.

It is the cutting of this very surface (a convex lens INSIDE a deep cavity) that is probably the most difficult aspect of his build by far. Instead, the collimating lens would need to be on the outside surface -- so that it would be convex on the "window" side where light exits, rather than flat. The idea of this build would to combine the light concentrating power of a large aspheric lens with a much smaller die image. You could probably get throw comparable to one of the mag aspherics, but in a much smaller, more manageable host.

A lens the size of the DEFT with a TIR system like this would be downright insane -- comparable to good HIDs for apparent throw -- but would also be insanely difficult to pull off...

2xTrinity said:

The corn oil/glass match there is as good as you can possibly get.

Click to expand...

Well not quite. The clear plastic that I use for the DEFT lenses has a refractive index of 1.49. Of course this is assuming the glass is 1.50 which it may not. Just depends on what kind of glass they used.

2xTrinity said:

would also be insanely difficult to pull off...

Click to expand...

Thats putting it mildly. I don't think some people realize how hard a task this really is.

Edit: I need to start reading the entire thread before jumping in!


Some nice out of the box thinking! I hope something materializes from it.

jashhash said:

Nah probably wont be able to beat DEFT But maybe come close. Do you think it would be possible to increase the throw of DEFT just by casting a layer of clear epoxy around the LED dome.

Click to expand...

There will be a mechanical stress issue and the dome may be broken off.

Also the aluminum surface around it presents problems. As it heats up, it expands differently than the epoxy and may separate. Epoxy and aluminum doesn't make extremely powerful bonds in most cases.

Yes. It happens all the time with Crees without even trying. The glass dome is very easy to knock off. The result is an LED with a reduced apparent die size which does help with throw but along with it comes a tint shift and reduction in lumen output somewhere in the 25-30% range.

Click to expand...

You know I was just thinking about that and a thought occurred to me. The dome of a cree LED tends to spread light out coming from the emitter. This is great for a flashlight using a reflector because that means a greater volume of light strikes the surface of the reflector. So when the dome is removed its easy to automatically assume that the flashlight is producing less light using the lux meter for comparison. I'm willing to bet that the amount of light produced is actually the same. Lux is a measure of beam throw and not of total light volume. The only way to verify if this is indeed true is by using an integrating sphere.

I'm not sure what the tint changes though. Perhaps its like passing white light through a prism and the result is a seperation of colors. Remember that when the dome is taken off all that's left is a layer of silicone with a jaggedy irregular shape. (update) from reading this post: http://www.candlepowerforums.com/vb/showthread.php?t=144141 it appears that the color tint, and output loss was a result of cleaning the silicone off the die and not by the removal of the dome. I'm guessing that through the washing process he may have damaged part of the die, removed a layer of phosphor, or changed the phosphors surface structure.

Flourinert is clear

Oznog said:

There will be a mechanical stress issue and the dome may be broken off.

Also the aluminum surface around it presents problems. As it heats up, it expands differently than the epoxy and may separate. Epoxy and aluminum doesn't make extremely powerful bonds in most cases.

Click to expand...

Yes there is a CTE mismatch but I don't think it matters in this case. For one it does not matter if the epoxy has a perfect grip on the aluminum. Even if it separates slightly it is not in the beam and will not affect light output. It really is a non-issue with my lights as the pills I make for the DEFT are already covered in epoxy and the clear can adhere to the white epoxy I use.
Also I am not in the least bit concerned about the dome popping off because the flexibility of the gel can absorb any stress caused by the CTE issues. We will find out though. I am making myself a pill that is encased in clear to test with tommorow.

I just wanted to chime in and mention that the 25-30% loss of light when the dome is removed was measured using devices very similar to an integrating sphere. Saabluster tested his using a ceiling bounce test, and I did similar tests using an integrating "box" that I made. Even though these devices may not be accurate when finding the exact lumen output, they are excellent when making comparative tests (before and after tests) since they do not vary any between tests.

Actually, I just did another unfortunate test on an LED that has a large beam angle: the MC-E. It was a pleasant neutral white emitter I had in a flashlight. I measured the lux when it was healthy, but after some "messing around" with different reflectors, I accidentally got the glass dome wedged into the opening on a plastic reflector, so when I pulled away the reflector, I took the dome with it. I tested the LED soon after because I was worried about losing total light output, and sure enough, I did lose about 30% of the total light. Now the neutral MC-E has about the same efficiency as the older cool white Luxeon K2 emitters. :mad:

So, I do not recommend removing the die unless you are really striving for throw. Even then, you may only see a 30% increase in trow when using a reflector (brighter, tighter spot), but then you find that the hotspot brightness is no brighter than the hotspot of an un-modified emitter (which is maybe slightly larger). So, all you see is a decrease in hotspot size and an overall decrease in total lumens.

I am interested in seeing how the large TIR optic would work. Didn't Inova use a very long, deep TIR optic in one of their lights? What if you start by mating that optic with the R2 XR-E or Sunlite emitter with oil? Just dont forget what saabluster said earlier: you are messing with the hard work of the researchers. You will have to put quite a lot of time into research to find the perfect materials, optical design, and mechanical design to seal this potentially slimy mess! Luckily, lots of flashlights come as being somewhat waterproof, so keeping the fluids in might be simple after all. :grin2:Now keeping the fluid away from the battery may be a different story!

I wish there could be more info on the "Sunlite 8W" LED emitter. I am wondering about the efficiency of the emitter compared to an overdriven XR-E. Dont forget the Cree XP-E (R2), the Lumileds TFFC Rebel (0100) and TFFC K2 (0200), and the Seoul P4 (U2) emitters are available for some decent over-driving. These emitters have the desirable lambertian beam pattern so you do not have to deal with the tricky optic and metal ring found on the XR-E. The domes on these emitters do magnify the die some, but they could be a better way to start. Actually, the dome on the Rebel can be removed (no bond wires -yay) so you can access the die directly (dry, no silicone goo found here!). Good luck all! Could this be continued in a more dedicated thread? Just curious...

Cheers,
-Tony

Thanks for the helpful insight, I wasn't aware that the light measurement was taken from a ceiling bounce. Now it's clear that the output is diminished but I'm still confused as to what's causing the efficiency loss.