Homemade 67mm glass based TIR optic!

[Ra]

Hi Guyzz,

I have a few things to show:

It all started a few years ago, when I machened down some of those widely aviable acryllic type of TIR reflectors, to put into small modified flashlights..
I still enjoy those flashlights very much.
Although TIR's have some disadvantages, as almost no fucusability (from spot to flood), very critical placement for uniform beampattern, they still give a very nice round spot with lots of throw and decent sidespill. And lumens efficiency can go up to 95% (only AR-coated glass based types!). When perfectly shaped, they can come very close to aspherical lenses when throw is the main goal (testing revealed TIR gave a 3-5% less candlepower output compared to an aspherical lens with the same diameter)

So, I wanted bigger, so about a year ago, I came up with the first glass based TIR optic, with a effectivity of 92% and a diameter of 30mm:

That one performed very well, so.. Going bigger, a few weeks ago, I again started to make a few TIR optics. The first, very successfull attempt was a 35mm diameter with a bigger central refraction chamber:

But a few day's ago, I finished this:

And the 67mm TIR in action: Beamspot..

So, I grabbed my drawing board, and made some prelimminary drawings of the torch for the 35mm TIR optic..:

I couldn't wait to share this with you.. Of cource there is more comming up: The 67mm TIR also wants to be in a torch like environment I think..

I'll keep you posted..

Regards,

Ra.

 

[jspeybro]

insteresting. care to share some info on how you design and make these lenses?

 

[ma_sha1]

Wow, yo should make a mold & make some large TIR with COC material.

 

[Al Combs]

She's a real beauty.

Is this the new one you mentioned made of BK7? Can't wait to see the finished light.

 

[LumenHound]

Looks great!

How many layers of anti-reflection coating do you plan on putting on the central refraction chamber and front of that 67mm TIR optic?

 

[Ra]

Looks great!

How many layers of anti-reflection coating do you plan on putting on the central refraction chamber and front of that 67mm TIR optic?

Well.. Enough to boost the performance another 7% !

And the inside of the refraction chamber will get an extra collimation lens.. (which yet needs to be made..)



Ra.

 

[Ra]

Wow, yo should make a mold & make some large TIR with COC material.

Hi ma_sha1,

Sorry, I only know my way in glass.. What is COC material, and of what material can I make a mold?

She's a real beauty.

Is this the new one you mentioned made of BK7? Can't wait to see the finished light.

Hi Al,

Yep, it is.. And the torch will be.. 'interesting..'



Regards,

ra.

 

[ma_sha1]

Hi ma_sha1,

Sorry, I only know my way in glass.. What is COC material, and of what material can I make a mold?

Regards,

ra.

COC is the best optical plastic, as clear as glass w/o polarization/ rainbow effect etc. I think the surefire TIR is made of COC.
http://en.wikipedia.org/wiki/Cyclic_Olefin_Copolymer

You can give dimentions & have a small run done by injection molding firm, the mold is made of of metal.

A 52mm Mag sized TIR for SST-50 would sell out quick.
The new surefire M3LT has a big TIR, but runs MCE.
A SST-50 TIR will eat it for lunch:devil:

 

[Walterk]

Looks great !
Can't wait to see what this magnificient shaped glass can do with your experiments with sanding down Leds and top-grade heat management ...

 

[Ra]

Looks great !
Can't wait to see what this magnificient shaped glass can do with your experiments with sanding down Leds and top-grade heat management ...

Well, if it doesn't work... It's also a nice candle stand !!


But it already performes well: 135,000 cp with a Cree XR-E R2 at 1.75 A (EZ1000 die) But that's without AR coatings and central lens

Note that I made a compromise: I wanted to be able to use this 67mm TIR for both the lambertian domed led's and the Cree's (smaller emittance angle)
So the Cree leaves a 23mm diameter central area unused, while with the lambertian domed led's there is virtually no lumensloss.
But when the central collimating lens is finished (17mm diameter), only about 190 square mm are not used by the Cree, thats about 5.5% of the total TIR surface.
lambertian domed led's use the entire TIR surface for throw.

If only I can get my hands on a few SST-90's (highest bin).... Anyone?


Regards,

Ra.

 

[PhotonFanatic]

.. .

If only I can get my hands on a few SST-90's (highest bin).... Anyone?


Regards,

Ra.

You can find some here. :devil:

 

[^Gurthang]

O....M.....G.......

Deft TIR!!!! *speechless*

sweetsweetsweet.

+10 on getting some COC castings made of both the 35mm and 67 mm.

*dreams of 35mm TIR in LF mini-turbo head*

 

[ICUDoc]

Great looking lenses, sungod. HOW on Earth do you make such things? I look forward to TIRs being available to mere mortals. When you say TIRs have a small candlepower deficit compared to aspherics, do you know what kind of efficiency difference there is also?? To rephrase that, do you know the lumens efficiency (to use your term) of an aspheric of equal diameter

 

[Ra]

Great looking lenses, sungod. HOW on Earth do you make such things? I look forward to TIRs being available to mere mortals. When you say TIRs have a small candlepower deficit compared to aspherics, do you know what kind of efficiency difference there is also?? To rephrase that, do you know the lumens efficiency (to use your term) of an aspheric of equal diameter

That highly depends on the setup: A single aspheric lens cannot grab the total emittance angle, so there you have quite some lumensloss. However, adding a pre collimator lens close to the led (DEFT) gives an important increase in lumens efficiency.
And although adding an extra (pre collimator) lens to the setup brings some losses (reflection, absorbtion..) The gain in lumens output is much more compared to those losses.

The side note is: This can easily be done with the Cree XR-E, because the 100 degrees emittance angle already is quite narrow (compared to lambertian domed leds) Those 100 degrees can easily be grabbed by a pre collimator lens.

The TIR can 'fold' around the led, capable of grabbing the full 170 degrees of emittance angle of the lambertian domed led's, and with high efficiency: Imagine the TIR putting out 95% of all emitted lumens from a SST-90..

Specsheets of lambertian domed led's like the Cree XP-G often speak of an emittance angle of 140-145 degrees, but in distribution graphs, you can clearly see that beond that, there still are some lumens to grab.. (up to 170 degrees) Not much, but lumens are lumens, and if I can, I'll grab them !!


So, in short:

With the Cree XR-E, an aspherical setup with a pre collimator can reach the same lumens efficiency as the TIR (all lenses must be AR-coated)
But with lambertian domed led's it's much, much harder for the aspherical setup to keep up with the TIR...(thats only on lumens output..)

For throw, the aspherical lens has the advantage of simplicity: It's hard, hard work to make a glass based TIR perfect enough to come within 3% of the candlepower figures of an aspheric lens with the same diameter.. (beleve me, I've been there..).. But then again,, I think making an aspherical lens is very hard for most as well..


Regards,

Ra.

 

[smokelaw1]

Imagine the TIR putting out 95% of all emitted lumens from a SST-90..

I've been imagining little else since I started reading this thread! I imaigne a P60 dropin with a single IMR18650 sending out a 8 or 10 degree beam of amazing intensity!!!

 

[^Gurthang]

Smoke,

Yep, can you say "light sabre"?? I can almost hear the *wsssszzzmmmm*....

 

[Michael Q]

Hello ra, do you use zemax, code V or other optical cad software to make your TIR optics?

 

[WmArnold1]

Hi Guyzz, I have a few things to show: ... So, I wanted bigger ...

Hello Ra; Why did you want bigger? Also, can you generalize which parameter controls the ratio of lumens between spot & spill??

Hopefully Yours, William

 

[bshanahan14rulz]

The more of the angle that the optic catches from the LED, the more lumens go to the hotspot.

The wider the optic, the more accurately it can collimate.

Isn't the optic a very basic function, its just hard to tool?

 

[WmArnold1]

Thanks for the fast reply, bshanahan14rulz!

Ra has said that a little spill is desired; something like 10% of the total lumens here. So, I assume that Ra's TIR's deliberately spill a little, even though I can't see how, looking at his pictures..

I'm not an optical engineer, but, Imho, one should be able to set the hot-spot cone angle and spill angle with the curvature of the face and a different outer curvature ring. Ymmv; I'm waiting to hear back from the guru: Ra!