Light collimation with an LED flashlight

As most of you already know, it is impossible to collimate all light from an LED flashlight since it is not an infinitely small point of light.
However, for the purpouse of this topic, I will attempt to explain how to “collimate” a light source as best as possible in order to obtain the maximum amount of throw.

First of all, we begin with a collimation template made by drawing shapes in sketchup, exporting as a transparent image, trimming away extra transparent pixels in photoshop, and then ending up with a PNG that looks like this:

Since the circle was made with 120mm diameter in sketchup, and the square was made with 203mm side length, when importing this image into Word you can set the image width and height to 8” (~203mm) and then the circle should be exactly 120mm when printed out, and the other objects at the same size and position and you set them in sketchup. Basically when we print this it should have all the same measurements as when made in the 2d drawing in Sketup.

The square size is not very important, you will see in a moment why that is the case. They just need to be centered at the dots, and the dots should be somewhat far apart but not near the edge of the optic.
The holes should be equally spaced away from the center, but not too close to the edge of the large circle. This is because the outer edge of most optics is not very exact, especially with lenses, and would result in us collimating incorrectly.

Two pieces of this template need to be printed. One gets small holes poked at the three dots, and then cut around the outside of the circle leaving a bit of white space. The smaller the holes are, the sharper the projected image will be, but you will need to be in a darker testing area and run the flashlight at higher power. I only used low mode, since the holes I used were about 1mm diameter.

When you turn the light on, the whole circle should be lit up by the optic, if you made the circle the right size.

Changing the flashlight focus will change the size of the light circle so don’t be too worried if it doesn’t match perfectly. The large circle is only there so we can center it on the flashlight head.
Here you can see I have about a 1mm gap all the way around between the light and the printed circle.

At a close distance you can see the 3 dots it projects. as you get farther away, these dots will turn into squares, due to the LED being a square.

The next step is to place the second template you printed out several meters away from the flashlight.
You want the square size you printed to match the square size projected by your flashlight, so that means maybe moving the template closer or farther away, or printing a different template with a different size of squares.
Ideally you want at least a few meters for accuracy.
PS- if you are collimating a reflector, instead of a lens, the square size will vary a lot depending on how far from the center you made your 3 holes.

On the left you can see the result irl, and on the right the 2d drawing to give a clear visual.
In this image, the flashlight is overfocused, resulting in squares which are too close to the center.
As you may have guessed, we want our squares to align exactly inside the black squares printed on the template.

Here you can see the exact opposite, the squares are spread too far apart.

Here all squares are focused at the same point. This is how you get a “clear die projection” when you don’t have a piece of paper in front of your flashlight.
All the squares coming from every point on your optic end up in the center, giving the highest lux at that point.
However, the following images will explain why this is NOT how you get the most throw.

As the light rays pass that perfectly focused point, they begin to spread out again, at a very fast rate.
Similar to what we saw in the second case here

What we actually want, is this:

Each “cone” of light is perfectly straight forward, with the center of the squares exactly at the dot.
This means that the light from the center of the LED is perfectly collimated (travelling in parallel lines) while the rest is + or - equal distance around it.
PS- since this is a lens, it has massive chromatic aberration, you can see the blue light separating from the red. This does not happen with reflectors. I just eyeballed the average center.

You may be asking “why straight forward”?
Well, if you take a look at the example where all the squares overlap, it is obvious why that would have the highest intensity.
However, we want this “focused” square to be very far away, not a few meters away, because otherwise it spreads out past those few meters.
To actually get irl the throw that you calculate with a luxmeter, this “point of highest intensity” aka “clear die projection” needs to be focused at infinity.
Here is an image that might help understand:

As the distance gets larger and larger, the projected squares grow in size.
Even they look very far apart at the beginning, just a few hundred or thousand meters away they will be almost completely overlapping.
This is because their side length increases to several meters but the distance between centers remains the same, just a few centimeters.
The result is large squares that are almost 100% overlapping, except for a few centimeters around the edge.
As the distance approaches infininty, the squares approach a “clear die projection” like here except much larger.

As you may have guessed, taking a lux measurement when the flashlight is focused like this is very difficult, because you need to be far away enough that ALL squares overlap, including ones from the outer edge of the optic.

The solution is to simply refocus the flashlight so that all the squares overlap at the luxmeter.
Since each individual square (or cone of light from a point on the optic) is affected by the inverse square law of light, focusing the flashlight to have them all overlap is just making things more convenient so that you don’t need to go 100+m away to take a measurement.
However, this should not be done for very close distances, since changing the focus of the flashlight will slightly change the chromatic aberration and size of the squares, so to get accurate numbers it is still recommended that you go a fair distance away to measure.

The final step is to try to imagine everything I explained but with an infinite number of these squares, coming from every point on your optic.
Hopefully this helped you understand how to get maximum throw out of your flashlight when focused properly, let me know if you have any questions :slight_smile:

3 Thanks

Thanks for the write-up.

I’ll have to give this method a go soon with a reflector based thrower, as I have a few Oslon Blacks inbound.

I usually focus my lights at 10m, as it’s a consistent environment that I have in my shed, without varying light conditions. It does the job, but I’ve always known that the focus distance is just that; 10m.

I’ll now use the method you describe, in the same 10m environment, to adjust focus on my throwers.


Thanks for this Enderman :smiley:


Cheers David

Wassup Starterman? Ooops!

I see you continue tinkering in your quest for a full spectrum white wide laser. Please add high colour rendering to it. :-D

By the way, saw you had a waiven collar in some sort of thread recently. Little question: Anyone tried attaching a couple high quality reflectors front to front?

I had to ask. :-)


You’re welcome :slight_smile:
Whenever I focus my flashlights at a spot near by, and then point the light at something farther away or into the sky, the beam diverges a lot and makes a big fuzzy spot which is definitely not optimal for throw.
Until now I’ve just focused my flashlights by adjusting them in the sky until the beam is as tight as possible, but I wanted to figure out a proper method to be sure I’m getting the longest throw :smiley:

Thanks :slight_smile:

Haha yup :slight_smile: unfortunately I had no idea the chromatic aberration was this bad, so I probably will not be using lenses for future throwers, just too many down sides.

What kind of reflector are you talking about? like a regular flashlight parabolic reflector?
If you put that in front of a wavien collar, no light will hit the reflector, since only 60 degrees of light exit the collar.

Thank you Enderman.


@ Enderman: I sure wish you are on the GT list too, so you can share with us what you can do with it, your way. :beer:

By the way, did you went 4S cell arrangement and classic FET @#$% with the BLF GT?

I'm loving the fully regulated 2S boost H2-C setup, the driver also proves quite capable of feeding all the meat an XHP35 requires safely and efficiently (see: Buck and Boost Drivers, Testing, Modding, and Discussion (Pic Heavy)).


Nope, I don’t really have that money to spend on something I will never use… :confused:
The only flashlights I own are either EDC, or superthrowers I custom buid.
The reason I would never use a GT is because it only has slightly over 1Mcd, and even with modding it would only go to about 2-3Mcd.
On top of that, it uses 18650 batteries while I prefer to use high capacity 3S lipos.

Also the BLF GT doesn’t have as much cooling as my lights do, so I wouldn’t be able to run LEDs at 6A+ while staying at low temps.
Finally, it’s a big and heavy light, and if I need to bring a backpack to carry some non-EDC light I might as well take my OptoFire with me which also fits in my backpack but gets 4.63Mcd.
Basically, I would need to buy 18650 batteries, an 18650 charger, and a $111 flashlight, convert all that to CAD and it would end up like $200+ which I would much rather spend on my next project :slight_smile:

I don’t have a BLF GT, the light used for this testing was my custom aspheric thrower “OptoFire”.

This is the whole point, your way to 2-3Mcd.

I’m at 4.63Mcd already, I need more.
Moooore :smiley:

Well, almost every one wants more actually, but my point is about the BLF GT specifically, not a light that looks like something like a shoulder-fired anti-craft missile launcher. :open_mouth:

At this point I’m willing to build a light that looks like a cake pan as long as it breaks records.
Someone else can mod the GT into a 2-3Mcd thrower, I just have slightly different personal interests :slight_smile:

Very well then, a record-breaking cake pan we go! :+1:

You will see it hopefully soon, maybe christmas :slight_smile:
3d model is almost complete, will start buying parts soon.

Followed a link to this topic.
Excellent method for tuning a thrower. :+1:

Would this method work for centering the led to the reflector and to adjust the height of the led in the reflector?

Nice work Enderman!

bookmarking for later use :wink:

Thanks, I knew people would find it useful :slight_smile:

Only for the height. In order to center it you would need to make the reflector perfectly parallel to the wall, and then have the template on the wall lined up with the reflector on both axis.

Thank you :slight_smile:

Too bad I’m almost indifferent towards optics (I really don’t know why) and throwers. This kind of stuff is not for everyone nor for every day use. I’m convinced that even years from now people are going to consult your posts. I think these things cost time. Kudos! :+1: :beer: