Compact optics required. What lens?

[mikeveal]

Help!

I'm trying to design a compact head torch.
From the front of the lens to the rear of the emitter, needs to be roughly 2". The slimmer the better.
The lamp is to be a thrower. I'm looking at using a P7 or MC-E.

I would like to use a lens to give a beam angle of around 2 degrees.

I have a number of questions:

1/ Where does the emitter need to be placed in relation to the focal length in order to obtain a focussed image of the die at infinity? (IIRC that's the definition of focal length, but schoolboy physics was a looong time ago, so just checking!)


2/ If I fix the focus of the torch at infinity, what will the beam look like when the illuminated object is closer than infinity.

3/ How do I calculate the image size, or beam angle given typical lens parameters? (diameter and focal length seem to be all you get.)

4/ I could use a mirror at 45 degrees to mount the LED at right angles to the lens, shortening the overall system fron to back. Can anyone recommend a supplier of mirrors? What problems might I encounter?

5/ Should I use a reflector on the LED to direct more light onto the lens? Will htis affect 2/ & 3/ ?

I'm sure the answers to these questions will raise yet more questions, but this will certainly help for now.

Many thanks for your time and patience.

 

[fareast]

deleted

 

[HarryN]

That is a serious technical challenge, but might be possible. If you use an LED with a single die, such as a rebel or k2, you might have a better chance.

2 degrees is very tight - if that is a real requirement, then you might try just lighting down a small tube, or array of them. I once shined an LED though a 2ft tube and it was amazingly narrow beamed, of course there was a lot of loss.

5 years ago I did the combo reflector + lens approach with a Lux V - 4 die package - similar optically to the LEDs you mentioned. It can be done, but the reflector needs to be a special design. What I found was that the optical models had a lot of trouble dealing with this, so I ended up doing a lot of it by experimentation. It still was not as narrow as you describe though, but I was attempting a 12mm dia setup, which is even harder. With 25mm dia, it might work.

A TIR optic would be one approach.

Think projector type optics.

Search the usual optics suppliers, carlco, fraen, ledil.

Send me a PM and we can discuss it more off line.

 

[mikeveal]

OK, I'll admit that I did not really want to go down the condensor / aspheric route. I was hoping to get away with a single aspheric lens.
I've looked at reflectors, and I'm not impressed with the output. Lots of light is wasted in spill and the beam divergence isn't that good.

Dropping to a single die would make the optics easier, but then light output is quartered. Since area is PI*r^2 then a MC-E with a divergence of 4 degrees would give the same illumination as a single die with a divergence of 2 degrees....
So dropping to a smaller die allows tighter optics, but they need to be tighter to have the same illumination.


I still need the qurestions above answered. Please ignore the 2 degree requirement. Pretend I didn't mention it. I'm looking for the formulae hinted at above to work out what lens I need... Then undoubtably when I look at what's available, I can start making design compromises.
TIA.

 

[Linger]

:welcome:Hello. Sounds like you've done a bit of thinking on this, but maybe not enough bench time yet.

-To get a small beam it is easiest to start with the smallest emission area.
-To get a tight focus you need a big deep reflector
-The only slim lights that are throwers are aspheric or TIR // and even then they miss a large amount of light. Observe the brighter aspherics will use larger lenses.
-As it sounds like you are trying for a collaminated beam, the focal distance (infinity or not) is moot.
-The beam will look like a lot of light, reguardless of the distance, I suspect you weren't serious with that infinity question.
-re: beam area, mc-e 4 degree divergence.
Sounds a bit like planning a trip to Mars before learning to walk.
Maybe you can spend $5 of a few reflectors and experiment a little bit. I suspect it will fast-track you past a huge number of conceptual hurdles.

Best,
Linger

 

[mikeveal]

To quote wikipedia:
"As mentioned above, a positive or converging lens in air will focus a collimated beam travelling along the lens axis to a spot (known as the focal point) at a distance f from the lens. Conversely, a point source of light placed at the focal point will be converted into a collimated beam by the lens. These two cases are examples of image formation in lenses. In the former case, an object at an infinite distance (as represented by a collimated beam of waves) is focused to an image at the focal point of the lens. In the latter, an object at the focal length distance from the lens is imaged at infinity. The plane perpendicular to the lens axis situated at a distance f from the lens is called the focal plane. (Note: In the figure below the image is actually larger than the object; this is a function of f and S1, described below)"


Infinity is considered to be any distance from the lens at which the rays of light from that point may be considered to be parallel when entering the lens. Or, put another way, when an object is close to the lens, you can focus an image projected by the lens onto that object, or you can focus it infront of the object, or you can focus it behind the object.
Whent the object is far enough away from the lens that you cannot focus an image projected by the lens on a point behind the object then that object is considered to be at infinity.


As you correctly point out, I'm not trying to focus a beam of light onto the big bang.




Still looking for equations to relate lens diameter and focal length to the size of the project image.

I'd have thought this would be basic stuff for you guys can no one help?

 

[ma_sha1]

Only Aspehrics can give you 2 degree beam.

Use top bin SST-50 & drive to 5Amp to get >1000 Lumens

Remove the doom to reduce the apparent emmiter size, you may lose
of 20% but still >800 lumens.

This might be as close to your original spec as reality allows, if you can live with a square sport.

 

[HarryN]

There is nothing basic about optics, at least for me.

It took a while and a lot of $s for the basics to sink in for me. In my particular case, about 3 years and $ 10 K, and I already have an engineering degree in another area.

The simple optical formulas are designed around simple, small areas. When applied to larger areas, then the real engineers bring out the simulation software. There is a reason people spend years in college getting degrees in optics engineering.

Pretty much all advanced LEDs today include an optical model for the simulators, and there are plenty of optical engineers for hire on linkedin or craigslist.

Try really hard to use a commercially available solution for a commerical or hobby project. If you really need a custom solution - seriously - hire an optical engineer. You might find that in spite of the MC-E being a very nice LED package, it could be the right or wrong package for your optical objectives.

 

[BillyNoMates]

HarryN makes a good point - it is not basic stuff - just look at the complexity of some of the lenses for a good SLR camera...

I'm not an expert in optics, but I do like to tinker a little bit with different lens solutions and I've been working on a (very) basic simulator to help me understand some of the issues - its a lot cheaper than buying different lenses just to find your idea was daft (and I've had plenty of those). I thought I would include some basic ray-trace outputs to help illustrate the some of the difficulties.....

First (as HarryN has already pointed out) the basic lens equation starts from the premise that the light is a point source - most lenses will be able to generate a 'reasonable' collimated beam if the point source is placed at the focal length of the lens.

As soon as one considers real world effects such as spherical aberration (this is just one non-ideal effect) and finite LED die size things go a bit wrong....

Consider an example - 1" diameter double convex "best form" lens, 100mm focal length - an example ray-trace is shown below:

As you can see, the point source on the left has been (almost) collimated by the lens. Infact, the beam is slightly converging as it has been focused about 3m distant from the lens (see later).

Now consider what happen when the rays start from a point that is offset vertically by 0.5mm (like a typical 1mm^2 LED die size) from the original point:

These rays have been plotted in magenta - notice how the collimated beam from this source travel in a different direction to the first set.

In the next graph I've expanded the x-axis to show the image focal point (horizontal position where the traces from both sources converge with themselves).

This lens set up would form an image of the LED die about 3.2m from the lens forming a beam with a divergence of 0.25degrees (0.5 degree cone angle) - sounds pretty good, doesn't? BUT..

Now consider the relationship between the source size (0.5mm), lens diameter (25mm), focal length (100mm) - not particularly useful for your application. NOTE: I also cheated a bit here - I only used a portion of the lens diameter which limits the marginal ray incident on the lens.
Faster lenses (shorter focal length wrt diameter) exhibit more spherical aberration and hence will form a more diverging beam. Also lenses with a smaller diameter will produce wider beams from a given source size.

Aspheric lenses can be used to counteract some of the effects of spherical aberration, which allows very short focal length lenses, but even these still suffer from the ratio of lens diameter to source size.

So, its not easy to recommend a specific lens for your application, but hopefully an understanding of the various trade offs will help you home in on an acceptable solution.

 

[mikeveal]

Harry, Billy, Ma_sha, and everyone else who's replied,

Many thanks for your time, it is appreciated.

I too have a postgraduate qualification in engineering, although mine is in electronics, not optics.

I do understand the importance of a pin point source of light if you're looking for a bean with narrow convergance. I'm assuming that this boils down to a ratio of the size of the emitter compared to the diameter of the lens. The greater the difference between these two, the less the effects of spherical abberation?

Looks like I'm just going to have to bit the bullet and experiment. Not what I wanted to do. It would have been much nicer to know what to look for in an aspheric before ordering.

I've ordered a bunch of aspherics from dealex (actually one of everything they have), and will have a play when they arrive.

Since I'm looking for a compact lens with a short focal length, the fresnel springs to mind.

1. Has anyone here played with fresnel lenses to project LED's?
2. Does anyone know of a source of fresnels that I could play with. Googling just turns up suppliers of Overhead projector lenses.

Thanks and Happy Christmas.

 

[Linger]

iirc getting an optical quality fresnel is too expensive. The economical options may have a projected image ringed with artifacts and distortions. It's OK if you're running a light-house but not so good for most flashoholics.

 

[Walterk]

And, how did you end up?

From what Ive heard plastic fresnel lenses work wel but have more light loss due to material and shape.

Ended up in this thread as I was looking for the beam calculations also.
I did some testing and interpolated from there but thats cumbersome.

My way, for easy reference I use the following (optimistic) guess:

Draw a line for diameter and focal length of the lens.
Draw the Led at focal point, showing the width of the die.
Draw a line from the edge of the die to the center of the plane surface of the lens.
Measure and double the angle between the last line and centerline and you have an indication for half the beamangle.