The LIGHTCANON - [flashlight]

[Enderman]

I thought my studio strobes looked like light cannons (Profoto B1's and D1's) -- nope, this is a REAL light cannon! Awesome work, and it makes me want to build one

thanks

still working on some improvements though

It's 90% finished now

[Exotath]

Great work this was fun to read.

[Enderman]

Great work this was fun to read.

I' glad you enjoyed it

[Enderman]

HOW TO CHOOSE A LENS - Part 1

 

A lot of different kinds of lenses are available, some cheap, some insanely expensive. Instead of having to do a ton of research, here's how to choose a good lens for your flashlight quickly and easily.

 

 

 

1) Aspheric. It must be aspheric, otherwise it will not focus properly due to spherical aberration.

 

 

2) Bigger diameter lens = bigger beam = beam is more visible

 

 

3) Higher quality = better focusing = longer range, but more expensive

Cheap lenses from Fasttech or other chinese manufacturers are less than $10 each, while professional companies like Edmund Optics or Thorlabs can cost hundreds or even thousands of dollars.

 

 

4) Less thickness (proportional to diameter)= farther focal length = smaller spot = longer throw distance

    BUT

    more light is lost = less intense beam = shorter throw distance

 

    More thickness (proportional to diameter)= shorter focal length = bigger spot = higher light efficiency

    BUT

    less throw distance

 

The trick is to find a good balance between small diameter thick lenses and large diameter thin lenses. The most commonly used lenses are about 70-100mm diameter and 20-40mm thick.

 

 

 

I will post pictures in part 2 of what the lenses you buy should look like!

And also how you can test a lens without buying it!

[Enderman]

HOW TO CHOOSE YOUR LENS - Part 2

 

 

 

So based on the points in part one, here are some visual examples of high curvature vs low curvature aspheric lenses.

 

Situation 1:

First let's see something like this:

Clearly the curve is very large, and typically 60-80mm aspheric lenses are look like this.

But why are these not good for long range? Well, below is a ray trace of light going through a lens similar to this.

*HINT* look at the light rays, some are going in but some are going out

 

And from a far distance, this is what happens to the beam:

This is with the beam as focused as possible. Moving the lens closer or farther away will both make the end light rays spread farther apart. You can call this the "sweet spot"

 

This happens because aspheric lenses, even though they fix almost all spherical aberration, they are not perfect, and light rays cannot be perfectly collimated.

 

 

 

Situation 2:

But now here is a larger diameter but flatter lens, let's see what happens.

As you can see, more light is actually lost around the outside of the lens. This is because larger flatter lenses have a longer focal length, so they need to be farther away from the LED.

But by being farther away, you can minimize the "spread" of rays that you saw in the previous ray trace image.

And when we look for far away, the beam is much tighter!

 

 

 

Now you may be thinking, "just make the lens bigger to use all that lost light" but if you use the program and try this yourself, you will notice one flaw.

Making the lens bigger will force the curvature to be less, making the focal point even farther, and you will need to move the lens farther away, losing the same amount of light as before.

If you try to counter this by increasing the curvature, you will end up with Situation 1 again, just with a larger lens. The light rays will spread exactly like Situation 1.

 

In summary, you need to choose whether to lose light and have a tighter beam, or to have a higher light efficiency and have more beam spread.

The only way this can be prevented is forcing the LED to output in a very tight angle, which is only possible with a Wavien collar, and that still only reduces the beam to 60 degrees.

(the ray trace images above have a 60 degree spread, I took into account the wavien collar. Without the collar you would have over a 120 degree spread, over 4x as much light loss (because pi*r^2))

 

 

 

Situation 3:

Here is the ray trace using the 100mm fastech lens (21.5mm thick)

I find this lens a good balance between light efficiency and beam tightness

Maybe in the future I will do some more accurate calculation on the area of light lost vs lens thickness, but for now I'm just relying on these ray-traced images.

 

 

 

 

YOU can experiment with lenses yourself by downloading this ray tracing program for free : http://arachnoid.com/OpticalRayTracer/

It might take a few days to learn, but it's very simple, and most of the lens "specs" can just be typed right into here

Any high quality lens manufacturer will have all the specs you need to make your simulation. Remember to use "parabolic"(aka aspheric) and not "spherical"!

 

 

 

 

TL;DR

-lens must be aspheric

-longer focal length = tighter beam /but/ more light loss

-shorter focal length = more beam spread /but/ higher light efficiency

-and you need to find a balance to suit your preference

 

 

 

I hope this helped! Let me know if you have any questions

[ChasingPluto]

 

 

4) Less thickness (proportional to diameter)= farther focal length = smaller spot = longer throw distance

    BUT

    more light is lost = less intense beam = shorter throw distance

 

    More thickness (proportional to diameter)= shorter focal length = bigger spot = higher light efficiency

    BUT

    less throw distance

 

 

I don't suppose there's a plug and play formula that we can use?

[Enderman]

I don't suppose there's a plug and play formula that we can use?

basically get a an aspheric lens with the diameter you want (usually the body you use will be the limiting factor)

 

then just choose between thinner lens (long focal length) for more throw or a thicker lens (shorter focal length) for more lumens

[Chrysolite]

Im intrigued  

So am I

[SSL]

lol, wouldn't have figured you as a flashlight nerd. You posted this shit on candle power forums or something yet?

[Enderman]

lol, wouldn't have figured you as a flashlight nerd. You posted this shit on candle power forums or something yet?

nah, I'm not going to make an account there just for one build log

i like this place better

[SansVarnic]

One question I have, is this a single lens project or are you going to use multiple lens as it would seem to me you would be able to focus you light and get better potential using 2 or 3 lens.

[Enderman]

One question I have, is this a single lens project or are you going to use multiple lens as it would seem to me you would be able to focus you light and get better potential using 2 or 3 lens.

Single lens

I don't see any kind of efficiency increase by using multiple lenses

 

due to the non-perfect collimation, using additional lenses would increase the spread of the beams

 

All 1+ Mlux flashlights I have seen on the internet are single lens

 

it may be possible to make a more light efficient dual lens system, but that would take far more advanced optical design, and much higher quality professional lenses

remember that a 0.1mm variation in a lens can mean a 10m variation when the light beam is going 500m+

[SansVarnic]

Single lens

I don't see any kind of efficiency increase by using multiple lenses

 

due to the non-perfect collimation, using additional lenses would increase the spread of the beams

 

All 1+ Mlux flashlights I have seen on the internet are single lens

 

it may be possible to make a more light efficient dual lens system, but that would take far more advanced optical design, and much higher quality professional lenses

remember that a 0.1mm variation in a lens can mean a 10m variation when the light beam is going 500m+

 

OK, I understand this.

But what I meant was maybe use a Biconcave lens after your aspheric lens with the collimation and increase your focus.

or do a parabolic reflector setup near the light source to pre-determine the light direction

 

Maybe I am thinking about this wrong...*shrugs* 

[Enderman]

OK, I understand this.

But what I meant was maybe use a Biconcave lens after your aspheric lens with the collimation and increase your focus.

or do a parabolic reflector setup near the light source to pre-determine the light direction

 

Maybe I am thinking about this wrong...*shrugs* 

I tried that with the ray tracer but it doesn't work very well because the LED is not a single point of light, it has an area

that means not all rays are coming from a single point, so more optical elements = more spread

 

I do have an idea of using a large parabolic reflector and an inverted LED to make a large diameter spotlight... but that will require more money and a more powerful LED system

Maybe in the future I will work on that

[SansVarnic]

@Enderman

 

This is what I was thinking

(did it in Paint so its not very good)

 

Either this or switch the two lens around around.

 

[Enderman]

@Enderman

 

This is what I was thinking

(did it in Paint so its not very good)

 

Either this or switch the two lens around around.

 

Light.jpg

well with that, just the parabolic reflector would already focus the beam, and that would be enough for a flashlight

 

idk what you're trying to do with the other two lenses, but the problem is that you're taking already collimated light from the reflector, then condensing it with the lend, then uncondensing it with the other lens, and you end up with the same thing as using just the reflector

[SansVarnic]

Oh, well I thought maybe I was overdoing it but I am glad I asked. The bi concave lens thing was an idea I cam across sometime ago (i cant reference, it was a passing thing)

Thanks for your input. 

[Enderman]

Oh, well I thought maybe I was overdoing it but I am glad I asked. The bi concave lens thing was an idea I cam across sometime ago (i cant reference, it was a passing thing)

Thanks for your input. 

that ray tracing software i mentioned previously works great to test things before actually spending money on anything

I wish I had discovered it sooner would have saved me some money

[SansVarnic]

that ray tracing software i mentioned previously works great to test things before actually spending money on anything

I wish I had discovered it sooner would have saved me some money

I will try it out.

[Enderman]

For future flashlight project:

 

 

Ok so I did some research and I found this:

A flashlight that uses "recoil led" technology

basically the LED is upside down so that all the light hits the reflector, instead of this:

 

This means much higher light efficiency = more intense beam+more lumens+ farther throw

 

I will do some ray traces, and plan to combine this "inverted LED" system with a reflector like this one:

http://www.edmundoptics.com/optics/optical-mirrors/focusing-concave-mirrors/large-parabolic-reflectors/53875/

 

this should give me near 90% light efficiency, no spill, and a 12" beam (I could also do 18 or 24", but that might be a bit too big to carry around)

 

And this flashlight would be liquid cooled, since the LED needs to be right in front of the reflector and having a large heatsink would block light.

[Peter Vaughan Truslow]

 

For future flashlight project:

 

 

Ok so I did some research and I found this:

A flashlight that uses "recoil led" technology

basically the LED is upside down so that all the light hits the reflector, instead of this:

 

 

And this flashlight would be liquid cooled, since the LED needs to be right in front of the reflector and having a large heatsink would block light.

 

how much current are you planning on driving the LED at, and what LED? I'd think that a heat pipe would block less light, but if you were driving an LED at insane currents, I suppose it'd be worth it (especially when you'd only be blocking a small fraction on a 12 inch beam.

[Enderman]

how much current are you planning on driving the LED at, and what LED? I'd think that a heat pipe would block less light, but if you were driving an LED at insane currents, I suppose it'd be worth it (especially when you'd only be blocking a small fraction on a 12 inch beam.

It will be a SBT-90 at about 8-10 amps

The reason I want a liquid cooler is because the block in just a small circle and then the tubing is all that needs to cross in front of the LED

 

If I were to use heatpipes there would be a lot less cooling because they would be over 6" long, and in this case moving the heat away is much faster with liquid

having the heatsink 6" away on a heatpipe is not as effective

[Peter Vaughan Truslow]

It will be a SBT-90 at about 8-10 amps

The reason I want a liquid cooler is because the block in just a small circle and then the tubing is all that needs to cross in front of the LED

 

If I were to use heatpipes there would be a lot less cooling because they would be over 6" long, and in this case moving the heat away is much faster with liquid

having the heatsink 6" away on a heatpipe is not as effective

that makes sense. a 120mm water cooler would have no trouble with that amount of power. it's kinda amazing just how well they perform with direct die contact like with a GPU. my H50 quiet (same as H55) cools a 170 watt card, with maybe around 150 watts to the core, with temps in the 50s.

would you be mounting it with or without the little star?

[Enderman]

that makes sense. a 120mm water cooler would have no trouble with that amount of power. it's kinda amazing just how well they perform with direct die contact like with a GPU. my H50 quiet (same as H55) cools a 170 watt card, with maybe around 150 watts to the core, with temps in the 50s.

would you be mounting it with or without the little star?

It will be with a copper star

I tried to find one that used a sinkpad or noctigon board but I couldn't find any with the LED premounted

 

The only thing now that im still unsure about is the LED+driver combo

Here the SST-90 only comes with a 9A driver and produces 2500 lumens: https://www.fasttech.com/products/1609/10010994/1940700-1-sst-90-30w-2500lm-3000k-warm-white-led-emitter

$25 is not bad, but I wanted something more

 

On the Luminus website the specs of the SST-90 show it can do 18A and 4000lm http://www.luminus.com/products/white.html

I dont remember it being 4k lumens last time I looked at this website a few months ago, they must have released a new bin

If I can get my hands on one of those and some 18A driver that would be awesome!

 

Large reflector spotlights really benefit from more lumens

[Peter Vaughan Truslow]

It will be with a copper star

I tried to find one that used a sinkpad or noctigon board but I couldn't find any with the LED premounted

 

The only thing now that im still unsure about is the LED+driver combo

Here the SST-90 only comes with a 9A driver and produces 2500 lumens: https://www.fasttech.com/products/1609/10010994/1940700-1-sst-90-30w-2500lm-3000k-warm-white-led-emitter

$25 is not bad, but I wanted something more

 

On the Luminus website the specs of the SST-90 show it can do 18A and 4000lm http://www.luminus.com/products/white.html

I dont remember it being 4k lumens last time I looked at this website a few months ago, they must have released a new bin

If I can get my hands on one of those and some 18A driver that would be awesome!

 

Large reflector spotlights really benefit from more lumens

 

I'm not absolutely sure, but I think you could parallel two 9A driver boards to get 18A, as they would adjust their own voltages to each drive 9A through the LED.

I'd not take my word for it, but with current drive you shouldn't run into any issues with evenly sharing current like you do with paralleling constant voltage sources.