Planning an aspheric scratch build

It helps to have a calculator that can re-use calculations and touch where you want to change values. A good android calc is MathsApp

Hmm. I guess thats similar to the small vis-basic app i once made, to run and re-run the calculations at the click of a moue i used for myself till i got used to this. (then a crackhead ran off with that laptop, also lost my reloading software i wrote for everyone on yet another site, ha ha)

same crackhead(s) got my adult air rifle, i had my working best prototype night vision gun mounted unit working on, to demonstrate it for people… so, if you ever see some crackhead with a “gun scope” with a fold-out screen on the side with analog imaging wires coming out of it, and an 8xAA NiMH battery pack wired up in the gun stock? GET them alone, and put a heavy glass asphere into a strong SOCK, and use it to beat them down senseless with it, ha ha… and rummage around in their house for my laptop, LMAO

maybe i’ll have to see about re-writing the visual basic program, i guess… it makes this fast and easy to see it…

PS - i tried the ray tracing program? it wants me top upgrade to WIN8, and t-h-a-t aint going to happen anytime soon, i know that. Maybe my other laptop (7) will run it, i dont know, next time i get it out and running, i’ll give it a try.

i just thought if someone used to ray tracing software could model this, it would cut the learning curve down a good bit less. You know, if anyone owns a well-working “pre-collimator” zoomie thrower?? this lets you enter the FL for both lens, and the air gap distance, and see how it works… if you built it you might already know the variables to enter….

…and you could see the “result FL” for the overall system get spit out.

Tests all done with a 100mm diameter, 75mm curvature radius, 19.1mm thick, plano-convex parabolic lens (closest to aspherical that I can do with this program)
Ray origin offset by 0.5mm to simulate the maximum angle of light coming from a ~1mm size LED.
Number of rays in test: 128 with a spread of 120 degrees (60 above and 60 below x axis) like a typical LED.
.
.
Single lens: placed at location of least divergence
Distance from LED: 124mm
Minimum spot diameter at 1km: 27.6m
Number of rays collected: 30

.
.
Double lens configuration A: placed as close together as possible
A1 distance from LED: 46mm
A2 distance from LED: 71.5mm
Minimum diameter at 1km: 41.2m
Number of rays collected: 63

.
.
Double lens configuration B: B1 moved 10mm back, B2 used to focus
B1 distance from LED: 36mm
B2 distance from LED: 94mm
Minimum diameter at 1km: 43.2m
Number of rays collected: 54

.
.
Double lens configuration C: C1 moved another 10mm back, C2 used to focus
C1 distance from LED: 26mm
C2 distance from LED: 110.5mm
Minimum diameter at 1km: 48.2m
Number of rays collected: 47

.
.
Single lens configuration placed in between A1 and A2 to simulate “joining” those two lenses together
Lens diameter: still 100mm
Distance from LED: 60mm
Lens curvature radius: 52.1mm
Minimum diameter at 1km: 180m
Number of rays collected: 61

.
.
Conclusions:
-With two identical lenses, the ideal configuration is having them close together (A is better than B or C).
-Two thin lenses are way better than one thick one, even though they will both collect the same amount of light (A is MUCH better than the test right above).
-A single lens (first case) still has better collimation than two lenses, but only collects a fraction of the light.

Additional points:
-The number of rays collected only represents a 2d slice, not the whole 3d column of light.
-The first lens in multi-stage tests (A, B and C) can be replaced by a smaller lens, such as in the image below:

Ok guys,

We need to start… Lenses source for multiple lenses flashlight system? Commercial available aspherical lenses can’t be used in such systems and even the one we can get have around 20% of performance play.

Camera lenses are very delicate one and far superior than classic aspheric lenses and for start I think we should start with them until we find reliable source for creating our own multiple lenses flashlight systems. So even cheap China ones could serve well… Or even better disassembled old M48(or any other old) camera lenses.

I listen you guys but my redneck head hurts from plenty of theory and numbers(sorry Seeds and other scientist members here; I hate math from the bottom of my hearth :slight_smile: ) cause my redneck brain rejects complexity and insists on simplicity.

But throw me some interesting idea and I’m hooked. Like precoolimator lenses at the days Vinz start using them and almost no one managed to crack that?… HA :slight_smile: Almost no one :slight_smile: I spend around 1000$ until I found right one that does not loose on performance one tiny bit and yet they enlarge de domed emitter projection for around 100. I saw Vinz builds claiming that he has 5 loss. I don’t have any loss with mine. That happened cause I was very intrigued by Vinz precoolimator as I am currently intrigued with multiple lens system.

I am not showing off I just want to say to Enderman and like-minded that there’s nothing that can’t be done… And again multiple lenses system is proven thing and it works.

Real aspherical flashlight fan would at least give a try to such system.

So scratch aspherical flashlight build? Give multiple lenses system a try…

I’m trying to deal with focal length requirements and effects. For example, a 3 lens system with lens 1 is closest to the emitter and lens 3 the farthest. What effect would lens 2 having a longer focal length than lens 3 have? Would there be benefit in using an achromatic lens after a biconvex lens 1?

Here is a page where you just type in the necessary values.

Mr enderman?

thank you from the bottom of my heart, for at least trying to model it. And my apologies for “how” i got you to do it. I know i got you to begin to try the traces in a very “base” and “crass” manner? By baiting you? But… it worked. I did the exact wrong thing, but, for a good reason and purpose. “Thanks”.

Maybe as people get bored, and get curious about the few precollimated lights that “work” someone might end up entering values that represent, for instance, the vinz light that pulls the trick off. Some of them are demonstrated to “work”, this should show in some ray trace somewhere, right?

let me keep going back up and reading and re-reading YOUR traces pictures, because (thank you) this is my first exposure to this power-tool. Maybe i can use the “online tool” he mentioned. I had the idea to start putting 3 or 4 thin lenses into the thing, and start changing their diameters and focal lengths and locations, trying to solve by “inspection”.

Like you? I would imagine that 2d is fine, nothing happening in 2d that isnt happening in 3d, as we’re modeling round objects.

Mr Enderman wins round #1… still, “bumblebees fly” and at one time, math of flight didnt yet “allow” a bumblebee to fly. Vinz’s pre-collimator flies, i SAW it happening in real life slowly over 8 hours of hand tuning, the emitter projection growing slowly brighter and larger.

Still? Round #1? is Enderman’s.

and thanks… and i hope this online tool works for me…

This multi-lens concept got me thinking about the possibility of a zoomie type light where the central lens is moved to change the focus similar to a zoom lens. This could be done using magnets mounted in a sliding ring around the head tube and in a ring around the movable lens.

oh. cool online calculator,just does the values though. still… neat.

one less thing…

You’re thinking about this wrong.
The way to make a spot larger or smaller is by changing the divergence of the beam.
For example a zoomie which can change the divergence from a few degrees to 60 degrees.
In that linked image all that’s changing is the beam width, it’s collimated in all three cases.
That’s not really useful in a flashlight because why would you want to change between a 5cm spot or a 10cm spot? 5cm -> 5m is a lot more useful
.
Also, it is currently impossible to make a beam narrower like you see in the third image.
Unless the light source is an infinitely small point, the light rays can never be collimated.
When you make the beam narrower, it diverges even more, causing it to become more uncollimated and form a cone of light.
.
This is why “beam expanders” exist, by making the beam a larger diameter, you can get less divergence, and closer to collimated light.

well? first off, i would be a lot happier if Mr Enderman would be a slightly less of a “good winner”. Quite frankly? Based on the “way” I enticed him in, baiting him in? He really “should” be enjoying himself just a TAD more than he is. When the numbers are in your favor at the moment? Enjoy it, i say.

I could care less about my feelings.

The more i stare at the ray tracings up there, and there are several… the more i am looking for any “shred” of anything that says anything positive.

Mr Enderman? While you are still in “gracious winner” mode? Is there any small situation where we saved at least one single “ray” ?

The endmunds optics paper, after you get tired of using “one single aspheric” the way we all traditionally do? They immediately describe using TWO of them, and if memory serves, the first is in regular position, the additional one added to the outside is “flipped around”, IE, the “flat side” of the collimator is to the LED, and the next one is flat side towards the target.

some danged aspheres are so thick it gets ridiculous? using a second one might put the first one so close to the LED source it wasnt even funny?

eh? unless the engineers at edmunds optics are writing instructional papers just for fun to have something to print…

Like this?

Some weird stuff is going on, looks like the spherical aberration you get when using spherical lenses instead of aspheric.

pictures small, i get 45 or 46 rays?

what happens when the lenses get closer to touching at the lumps?

@Enderman The portion of the linked image I was interested in was the central lens and was considering the possibility that divergence might be controlled by changing the position of a lens or pair of lenses relative to two fixed lenses. The motivation for the thought process is a zoomie with no twisting or sliding segment, just a sliding ring or other magnetically linked mechanism controlling lens positions. I’m new to this and still building a mental model of how different lens types and characteristics interact.

did you look at the edmunds optics white paper? if only out of curiosity? On “page 1” (part 1) they go over, basically, our entire basic site strategy. Which is one lens. They describe the LED source in terms of theta half angles, and the lens in terms of “NA” which is numerical aperture.

in terms of a camera, “aperture” makes sense to me. the numerical aperture of the beam, in this site wide strategy,is basically the sine of the half angle of the LED emission. They use an example of the LED with a plus/minus 50 degrees emission of half intensity. sounds fairly representative to me.

you get NAbeam = sin(100/2) = sin (50) = .77

I mean, THATS not hard, the sine of the half angle of the LED emission? is the “numerical aperture”. Any fool (like me?) with a TI-30 or the budget equivalent, can play along at home. If what you wan to illuminate is close enough to that beam? you dont need a LENS, but, we want to go further than a few inches.

you get a lens, which should sound familiar to any zoomie enthusiast. First, you want the NA of the lens to be greater or equal to the NA of the LED emission. the theta angle stated in radians, is the diameter of the source divided by the focal length of the lens.

namely, Theta prime(radians) = Ds / FL.

sounds a lot more complicated than it really is, radians convert directly to regular angles directly. Now, if you have born this out THUS far, they state at that point:

“This identity shows that a small beam
divergence may be achieved by either
reducing the source size or by increasing
the focal length of the collimating lens.”

and REALLY all they have done up to now? is state mathematically things we all already talk about. We already KNOW we would like a small sized LED, and we want the emission angle to be as small as we like. The xpg series being a “thrower”? Is simply described mathematically. We still havent moved off of anything we all dont already know, and that we all already talk about. The math hasnt gotten bad. Anyone awake in high school angebra and trig is still alive.

we cant GET a better LED yet, so, we increase the focal length of the lens. Gradeschool BLf, this is what we explicitly “do”.

They immediately state:

“As
the beam divergence of the light source
is reduced, the collimated beam diameter
increases proportionately. The collimated
beam diameter (Dc) right after the collimat-
ing lens can be estimated as follows:”

Dc = 2 * NAbeam * Fcol

which instead of us standing around and saying “wow! cool pencil beam!” or “hey, nice even flood!” ?? they mathematically state the diameter of the light beam. the number two, times the numerical aperture of the led, times the focal length of the lens? yields the beam diameter.

I can multiply on a calculator, i dont feel left out in the cold yet.

okay, so heres the “money shot”… they state our entire problem, still on page one/part one? as thus:

“As the focal length
of the lens is
increased while the
lens diameter stays
constant, there is a
tradeoff between
maximizing light
throughput and
minimizing beam
divergence”

THERE, in the middle of “page 1, part 1” our entire site strategy? and indeed,the entire site strategy of my night vision site? (IR zoomies, same thing) is summed up with MATH. you want better results? Theres a tradeoff, and we all know it. Hey, i’m a “redneck”, I l-i-k-e a “bigger hammer” theory, lol.

I have no problem with putting more amps thru the LED and doing better heatsinking. Its a bigger hammer. I like using the xpg or the best led, its a better hammer. I understand that using a larger diameter lens is a bigger hammer, and the redneck in me “groks” this perfectly. But, the paradox here is stated mathematically.

NOW, at this point? my night vision site? had us “fanning out” and infiltrating the flashlight sites, lol. We were running into the same “wall” as you guys here. longer focal length, wider lens… and you run into whats practical, and you run into the paradox.

==

in summation? the math hasnt gotten crazy. nothing beyond a few key presses on a TI-30. No gigantic equations. Youre doing ray traces, i assume YOU are still with me, right? The only thing i LIKE here? couple of things…

the entire “problem” is stated mathematically. No hunches, no “its all about the throw, dude” talk. They DO STATE your “diameter = throw” argument? They state it in a way that means something concrete.

Now, i know what you will say. “big deal, same problem we have, stated a different way” which is true, but This is page one, part one. They dont get off of PAGE ONE of the engineering paper? without the next best strategy…

i swear to GOD himself, theres no worse math than what we encountered here. Real simple. They tell us, the very next step. Will you permit me this indulgence? its just the next paragraph or two, and its easy to follow intructions and an equation no harder than what we saw here. easy.

well? I hear nothing but “crickets chirping”. I might well be once again talking to the walls, which annoys my cat. But, just in case…

quick recap? our paradox has been stated in mathematical concrete terms. we KNOW the pre-collimated lights exist, examples of them have been shown. I myself have built two of them, but in infrared. a long process of “hand tuning” 4 lenses interacting with each other? working invisible infrared? is “maddening” to say the least. notwithstanding, the only way io can MAKE them? is both times with… the lenses from three gun scopes all tore apart, and also one spotting scope not made anymore. and the long hand tuning process.

vinz made them, the violin maker has one… bigfoot exists. And once you have SEEN the emitter get bigger AND brighter,instead of bigger and more dim like it should? you are HOOKED like CRACK.

i dont think these people are “calculating” it? i THINK they did what i did… noticed it during long “hand tuning” and capitalized on it. BUT, thats my guess. Either way? none of them are forthcoming, exactly “what” to buy and “what” to exactly DO to replicate the results.

this engineering paper i am referencing? which no one will read? immediately moves on to “step two” which they entitled “using lens PAIRS”…

QUOTE
2 Aspheric lens pairs
A second aspheric lens can be added in front of the collimating lens to focus the
light over a desired area. The illuminated spot diameter at the focal plane (FP) is
given by
ENDQUOTE

SpotDiameter = Ds * (F-foc / F-col)

Ds = diameter of the source LED
F-foc = focal length of the focusing lens
F-col = focal length of the collimating LED

===

they immediately follow this with I_N_S_T_R_U_C_T_I_O_N_S of what to do, and to select lens pairs! to wit:

By selecting the focal length of the focus-
ing lens to be longer than that of the
collimating lens, it is possible to reduce
the numerical aperture to below that of
the source (NAbeam). For distances greater than f focus, the full
beam divergence (θ’) in radians becomes:

THETAprime = 2 * NAbeam * (F-col / F-foc)

THETAprime = full beam divergence (expressed in radians)
2 = “the number two”
NAbeam = numerical aperture of the beam
F-col = focal length of the collimating lens (inner lens)
F-foc = focal length of the focusing lens (outer lens)

==

Thats it. Nothing else! they merely go on to say that the “limitation” of this? is that the “source structure” of the light source will be projected out, which is a way of saying “you will SEE the emitter” which we all see anyways, big deal.

the whole rest of the paper? has nothing for us, its a bunch of crap to “blur” the light so you cant see the emitter, crap like light pipes and microlens arrays, nothing concerning me.

NOW… this description? they are DESCRIBING the “vinz” pre collimated light! this is the perfect description of the “violin maker build” referenced above. THIS is what i saw happen with “hand tuning” 4 lenses! why did i see it? i know now… my 4 lenses were actually 2 lenses, really… my “rear 2 lenses” were interacting to be the collimating lens, and the “front 2 lenses” were interacting to be the focusing lens!

as i hand tuned the 4 lenses around? i happened onto the “trend” and was able to chase it and maximize it… because my 2 lenses, were each made out of 2-lens-compound lenses with air gaps? moving the lenses around was like selecting different lenses! i didnt have to try hundreds of lens combinations, i had them all available by adjusting airgaps!

do you see?

===

so? for the final ray trace? i would suggest…

a lens with higher focal length, on the outside… and a lens with lower focal length, on the inside in front of the LED.

naturally, i am unsure exactly what the values of the FL’s should be…nor exactly the distances, but…

…isnt that what fooling around with the ray-trace would show? at some point, you would notice “rays” being captured in excess of a single lens??

this isnt “conjecture” on the internet, this is an engineering paper, the exact sort of paper “dr jones” would follow to make something in his lab for whatever company he works for…

its “how” to make the precollimated light. we just need to fiddle with the equation values, and the lens choices. ray trace should show this??

===

lets be “prometheus”. lets steal FIRE from the GODS ? and give it to MAN… lol

you see, in hindsight? everything makes sense. i rmember reading about one of the precollimated lights someone on here debuted. i dont know if it was djozz or vinz… someone did…

anyways, i rememebr the long time they spent getting JUST the right little lens above the led… then the amount of time it took to get it JUST at the right height…

also? the reason we have to trial hundreds of lairs of lenses? is because we are trying to “hit” a RATIO of the 2 lenses focal lengths… change one lens? the other needs to change too. you cant move the lnses from the optimum spot either, at least not the collimating lens.

its not the “exact lenses” per se? its the RATIO of the focal lengths that is at issue!

i have ALL the ingredients for this dish right here, in one place… help me!

i mean, we have spent ENOUGH YEARS now, in the flashlight world… on “first half of page one”… why cant we “bravely” do the rest of “page one”? you know?

we want these lights… this is how.

i know, this is the work of “doctor hoozits”, right? well, he’s not here to help us… i found the information and its not that bad…

:slight_smile: Yes you can’t really work with IR light and test bunch of lenses… Even cigarette smoke or coffee will not help here…
That is real pain cause even NV googles will not help to determine weather you are on good way or not, but lux meter can help a lot…
In my caste to crack Vinz precoolimator lenses I ordered a bunch of lenses(smaller 10mm diameter lenses of course) from all kinds of sources and since Vinz gave few important hints and one of the most useful hints is that “lenses need’s to be placed around 0,7mm from emitter and it is very tricky to place them”. Rest I did was hand fitting of whole bunch of lenses and lux meter testing each of them(o rings can help here) and ta dam! I found them. How I knew it was the good one?
If I for example know that B158 has around 300kcd with classic de-domed emitter and after sticking precoolimator does same thing but this time with larger main beam and even improved tint(yes precoolimator improves a tint shift) you know that you have found the winner.

There are certain lenses that can for example enlarge XP-G2 for more than 500% (yes I am serious here), Crazy :heart_eyes: I was shocked to see monster XPG2 Projection with 50mm aspheric on wall but my happiness disappeared after putting that on lux meter which showed around 40% less lux performance and not only that very poor flood mode after defocusing flashlight. And why that happens that is thing that only you doctors know :slight_smile:

So that was base for me to start and if anyone will want to try that it has good base to start from this post. I respect Vinz for that idea and I am really greatfull even for such info and I fully understand why he did not want to help me with lenses selection (this is is his hard work and something that distinct him from other modders cause it really ain’t problem to de dome and place emitter into aspherical flashlight but it is really big problem if you want bigger main beam projection without any lux loss).

But I don’t consider precoolimator lenses as double lens system(although it is) cause when I stick them the only thing I get is around 100%(possible little less) bigger die projection in full focus mode without loosing any lux.
But on the other hand I loose on maximum flood mode… Flood is around 20-30% narrower than usual non pre coolimated de domed emitter. But also worth to mention that flood mode(to the point you can defocus it of course) looks more intense then on regular de domed emitter.
So if someone wants impressive flood mode or more lumens out of flood mode that ain’t going to happen with precoolimator lenses and single aspheric lens above it.

Real success would be to have a smaller multilens aspherical system that could without any additional precoolimator lenses on emitter perform same(or even better) than his bigger single aspherical brothers and if that with real experiments is not achievable(although we now that is achievable since Sunnranger works with IR light and even with white light - but it looses on performance with white light) than there is no point of multilens system except for fun, and we are stuck with “the bigger they are the further they throw”.

I will do a new set of ray traces later with different focal lengths on each lens, going from short+long all the way to long+short and we will see what happens.
Let’s see if there’s an ideal ratio :slight_smile:

well, hold on mr enderman.

i am trying to not send you “willy nilly” ray tracing everything under the sun. i view you doing this, as a “finite resource”, and i figure i only have so many “ray traces” and so many “manhours” before you give me the finger, lol.

PS - before we start again? we probably should select an LED we would want to use for a “thrower”. Would this be the xpg2 ? I noticed your ray traces all used the “1mm” source size example, this says “xpg” to me.

also? we might as well start the “model”, from the paper. Carry it through “a best build practice” single lens zoomie. If what it predicts, even closely resembles what we “know” from real life? Then it will make easy sense to move forwards.

rendering the fat mathematically? I have been espousing the “value” of the f-number. While you have been espousing the value of “diameter increases throw”. Right off the bat? “we are both right”.

step 1, in this whole mess? starts at the LED itself. its emission half angle. you used plus minus 60 degs. Thats an initial theta of 120 degrees. THIS is step one, we establish the NA or Numerical Aperture of our LED. The sin of 60 degs, is .866

Thats the “cone” of where all the best amount of light emitting from the LED is coming from. Now, right off the BAT, to model a single lens zoomie? we want THAT cone to match the NA of the single lens, quite naturally. We can best express THAT angle, the numerical aperture of the single lens? With the humble F-number. Because it relates the diameter of the lens to the focal length? Which chief soh-cah-toa will produce an angle for me. We need that angle or greater. Stated otherwise? we need “a certain F-number, or lower” or else right form the get go, we are wasting light.

let me see… setting up the chief soh-cah-toa… if the half angle is 60 degs? what f-number of 100mm diameter lens will have this emission angle matched?

the triangle sets up with theta=60… half the diameter is the “OPPosite” side. i find a hypotenuse of 57.737, useless side by itself? but, it sets up the “adjacent” which is actually the focal length of that triangle… cos of theta 60 is .5
and… i am getting for a 100mm diameter lens, the theta 60 yields a focal length of… good sweet christ… 28.9mm

(check my math, but…) 60 degrees half angle? is BIG… a 100mm lens has to have appromimately focal length of 28,9mm focal length to contain that angle, if you want the numerical aperture of the lens to match he emission half angle of the LED of 60 as you stated as “typical”.

that means, any single lens zoomie? to have a chance? has to have a F number of .289

that is a TALL ORDER!!!

fortunately though? we can create F-numbers out of thin air? we already know from the compound lens formula? that any 2 lens touching airgap zero, will yield ONE lens of exactly half the focal length… so, every time we add a second lens? we are cutting the F number in half…

this is a RIDICULOUS figure! (make a nice flooder though, lol)

now then, are you SURE about that half angle of 60 degrees? as typical for the xpg series?

i would be a lot more confortable with something more “sane” for an f-number requirement of the lens… like F 1.0 ?? geezus…