Talk about future projects and donation topic

BTW, Enderman, do you have an idea how much would a top-notch collar cost in volume?

isn’t it just a reflector with a very large hole and a specific wider curve?
I don’t see why such thing would be very much more expensive then a reflector

OK, did some tweaks:

  • added a switch
  • added threads for battery insertion
  • enlarged zooming threads a bit
  • switched lens to 40 mm EFL one. This is a big change. It freed a lot of space all over the light.
    Not done: I should redo the collars as now one blocks the other. This would lead to redoing the head. Outside dimensions wouldn’t change much, so I skipped this step for now.

A note about collars vs. other techniques:
Collar is the only optics that actually improves beam intensity over a simple lens. Otherwise, to get the same effect you would have to increase lens diameter. The zoomie above has a 50 mm lens and does 400+ kcd with SST-40. If it was collarless we’d have to scale it up to 70 mm to achieve the same effect. Body length would increase by nearly 20 mm as well.

I haven’t found any 14500 or smaller flashlight that can do 250m throw for 4h on max like my P5r.2 can do.
If you put a more powerful LED in it you can probably get 500+m throw out of that tiny thing :slight_smile:

It’s not really a beam until it exits the light, the precollimator just reduces the output angle from the LED.
The spot created by the flashlight will always be larger when using a precollimator.
It collects more lumens, and makes the spot larger, essentially leaving the same intensity.

There are other things such as glass transmission and surface quality that will affect the performance though, so you can probably expect less lux when using a precollimator.

Adding a precollimator essentially reduces the focal length of the dual-lens system (this can be mathematically calculated), and as you probably already know a shorter focal length will mean more lumens collected and a bigger spot.

It’s called a cold mirror, and they typically reflect 99% of light in the visible range.
Some people were discussing in PMs about getting some aluminum ones machined.
A typical aluminum reflector like you see in flashlights is 75% reflective, not 90%.
If you want to get 90% you will need an electroformed reflector, which is also expensive, especially since nobody produces them at the size that would work for a collar (yes i’ve checked).
A custom job will cost several thousand dollars because they need to produce a mandrel for it.

No idea, it would depend on what manufacturer you go to, and how many you want produced.
A single one will cost about $1k.

Thanks, I need to get some basic optical design knowledge…

Where did you get this number?

> where …

Lots of instances: typical aluminum reflector like you see in flashlights is 75% reflective - Google Search

click/tools/verbatim to limit the search results

I checked all the first links.
I did not find any statements regarding this except an older post from Enderman where he claims the same thing because he thinks that normal flashlight reflectors are polished. From what I can tell the majority are coated (after first being polished). Every standard flashlight reflector in good lights has a coating of vapor diposited aluminium. Even the plastic maglite reflectors have this in some form (maybe not aluminium, not sure). It’s nothing special and not very expensive.
.
I have never seen anyone do actual measurements with a high degree of certainty and low possible error which prove that standard flashlight reflectors only reflect 75%, never.
.
It also has nothing to do with electroforming. Electroforming is the process of creating the actual shape of the reflector. This shape (made of nickel) can then be coated using different methods. It’s mainly needed for higher accuracy!
.
Sere here for an interview of the ceo of 4Sevens where he explains how they produce their reflectors: How LED Flashlights Are Designed And Manufactured
4sevens offers a 30$ regulated EDC light with bluetooth functionality and such a reflector. It can’t really be that expsenive to use such coating if they can do it at this price point!
.
Here post from Ra (optics engineer) also agrees with this. He says that aluminium reflectors reflect 88% of the light (the total efficiency he talks about is referring to how much light actually goes into the hotspot compared to a TIR lens, which he prefers).
.
Or here:

First 2 results I see include the 75% figure. But Google does not show identical search results to everyone for the same query.

More turns up with a focused search:

(I try not to get caught up in arguing over there, some of those threads get a bit snarky)

http://candlepowerforums.com/vb/showthread.php?106790-bLu-vs-tLu-IS-confirms-65-conversion-factor

The reason it seems like less than 25% loss when you measure the flashlight lumens is because not all of the light is actually hitting the reflector, about 1/4 to 1/2 of the light exits the flashlight as spill, unaffected by reflector losses.

Maybe newer technology has changed this since then, like with the new nitecore reflectors and stuff like that.
It’s not going to be anywhere close to 90% though, you need electroforming for that.
Maybe like 80-85% at best.

The only real testing being posted so far.
But even their test is not perfect. They didn’t actually test the bulb without the reflector. Bulbs have manufacturing tolerances like everything else. Also, they are using a potted bulb. The base that it is pottet into obviously absorbs some of the light that the bulb reflects towards it’s legs.

So in general it doesn’t make much sense to compare the values the way they did. The should have tested the potted bulb without the reflector first.

Their 65% factor is probably accurate, but only for flashlights with bulbs and uncoated lenses! The lumens of lightbulbs are reduced more compared to LEDs.

Concerning Nitecore:
I would be rather suprised if they actually did something special. To me it sounds like they invented nice marketing terms for the same processes that 4sevens uses (aluminium pvd coating and afterwards a clear coat to prevent oxidation).

EDIT: I found another test. He tested a regulated, modded LED maglite with different reflectors, lenses, with and without bezel. He also tested for a baseline value. His numbers seem a bit off though. Borofloat glass should be around 94%, he measured 98.
The problem here is that one really needs to use a rather accurate luxmeter-sphere (if one doesn’t have access to a real calibrated IS) to get meaningful results. Otherwise you just never know if you measured something useful.

Point is, the reflectance of the reflector is not anywhere near 90%, it’s closer to 80 I would guess.
Electroforming doesn’t only make a more accurate shape, but a lower surface roughness also increases the reflectivity.
Absorption and scatter depend on how accurate the surface is, to fractions of a wavelength of light.

No matter what the actual values are, I think this topic deserves a more in depth Look.

Agree.

So when you guys are talking about the efficiency or reflectance of a reflector your talking about how much light hits the reflector and how much goes out the front, missing the reflector.

I thought you were talking about the efficiency of the reflector coating which has to be like 97+%.

No, we’re talking about the reflectivity of the reflector surface.

It is 75-85% for the typical aluminum flashlight reflector.

A precision electroformed reflector with aluminum coating is 92%.
An extremely expensive and delicate silver coating is 97%.
A dielectric stack or cold mirror is even more expensive and approaches 99%.

Yikes! I see a mirror looking finish and just assume it’s more than 85%.

Nope, not even close :frowning:
Even the silver coating was unaffordable for me for my custom reflectors I’m ordering.
In order to not react with oxygen it also needs special protection coatings.
The problem with cold mirrors is that they need to be made of glass, so instead of $400 per reflector I would probably be paying $4000.

Another set of zoomie drawings.
First, in the zoomies I drew before I confused some optical parameters. Most importantly focal length with back focal length. Zomming action needs to be much shorter than I anticipated.
Second, I actually used a real lens in the drawings. It was very expensive, but I hoped there wouldn’t be a big difference if a cheaper one was used. I switched to a cheaper one now (though still not cheap, €49 when buying 1 piece).

As you can see, threads are way longer now and I stopped worrying about thermal transfer. The whole light is slightly shorter too.

Since focal lengths can be really short, I wanted to investigate a zoomie suggested before: where the precollimator is the only moving part.
But I haven’t acquired the skills to draw one yet….so instead I drew a non-zooming aspheric with even stronger lens

Side switch, thinner battery tube, ~13 mm longer than the zoomie above. With a pre-collimator it could possibly be even slightly shorter. As it is, even taking into account my drawing’s sloppiness, it should be shorter than B158.