Well. You are 100% correct. I got a light with a frosty drilled optic. I assumed it was a 10508, polished it out, and it was still very floody. Upon closer inspection with a known 10508/10509, I definitely had a drilled 10509.
Thanks so much for the information! I thought I knew about Carclo optics, apparently I don’t…
The discussion here seems to revolve mostly around domed emitters like a 519a. What’s the difference that the 10621 is likely to make over the 10622 for a dedomed 519a?
I ordered a 10621 from jlhawaii.com for my D4K with stated emitters. It seems to have focused the hotspot a good bit more, and seems to have made the tint a bit rosier. There seems to be some artifacts in the beam, but not enough to really bother me.
I tend to prefer throw over flood, with some usable spill, and from what I understand, dedoming an emitter like the 519a is supposed to narrow the beam some, then putting the 10621 on top of it should narrow it more.
As you observed, the 10621 gives a rosier hotspot tint, which holds for domed and domeless emitters.
There’s a difference in beam profile: the 10622 has a small hotspot surrounded by a large corona, while the 10621 has a larger hotspot and essentially no corona. The tradeoff is having more artifacts around the edge of the hotspot and in the spill, since domeless emitters have less rounded boundaries, and the 10621 has sharp edges within each of the four optics.
The actual throw probably won’t be much different, the 10621 might win out marginally because it has a larger central convex lens that takes advantage of the fact that the angular distribution of real emitters isn’t perfectly Lambertian but more intense (and rosier) in the direction perpendicular to the LED’s surface. It’s thus also the reason why the hotspot is rosier.
Necro bump - running into this thread searching for something else. I had a chuckle looking at your detailed (and correct) description. Here is a picture and measurements of 2 LEDs, 219b 4500k and E21a 4500k in D4V2.
One thing I’d like to add is there’s a change in tint of the “corona” or whatever that yellow ring in E21a beam shot is considered. And how it magically disappeared with 10623. I believe this is the reason Hank shipped 219b with 10622, but E21a with 10623. I have observed the same with Convoy S21D with 2 different TIR: 10 Clear vs 10 Frost. 10 Clear has quite a nasty yellow ring that 10 Frost improves.
Thank you for sharing these beamshots and pointing this out!
I have not considered in my original writeup emitters whose phosphor coverage extends past the die, such as the XPG3 and the E21A. These LEDs have a yellow outer ring surrounding the die that require diffusive optics to smooth out–any sort of smooth optic, such as the 10621/22, would retain the yellow artifacts. The 10623 on the other hand is excellent for smoothing out artifacts and tint shift.
You’re welcome. I too feel great that your post “validifies” my photography.
Side question regarding E21a. I just measured D4V2 with E21a (~870 lm) vs with 219b (1600 lm). I couldn’t believe such low E21a output but it seems c/w ZeroAir review. FWIW to me the 2 beam colors look identical in real life also with E21a max output much dimmer, so my feeling is E21a is similar to 219b, but with added artifact and lower output. Is that too harsh of an opinion?
It is always satisfying to me when theoretic predictions are consistent with empirical observations!
Your opinion is completely justified. In color rendering, the E21A has nothing to offer over other blue-pumped R9080 LEDs such as 519A, 219B, B35AM, etc. But the E21A has very low output, and the beam profile is so bad that clear optics cannot be used.
To get any significant improvement in color rendering over blue-pumped R9080 LEDs such as 519A, the dip in the cyan region of the spectrum must be addressed, and given current phosphor technology this seems unsolvable due to its proximity to the pump wavelength. Thus the only real way to get better color rendering is violet-pumped LEDs (such as Optisolis, SunLike, etc.), or revert to incandescent.
Does this apply to 9xE17A (DA1K) too? Is the 10392 ok or do you need the 10393 to smooth out artifacts?
I don’t know about mechanical compatibility of these optics, but for getting a quality beam out of 9xE17A, you’ll need as much diffusion as you can get, so I’d prefer 10393 over 10392. The weaker frosting won’t do anything for this choice of emitter array.
From my observations, frosted surface loses a lot of throw. I was a fan of frosted TIR but I gradually started to prefer clear surface. I can accept a bit of beam artefact since it only affects white wall hunting, and in real world use I far prefer more candela on the target.
That would also be my preference for most emitters, but 9xE17A really cannot do with anything less than the strongest diffusion (for both beam profile and tint consistency), and if one were to choose this configuration in the first place, hopefully they would be well aware of how much throw they forfeit over, say, a single FFL emitter with the equivalent power. TBH I would never make this trade, since the E17/21A does not offer any better CRI than the other more powerful and better-behaved R9080 emitters, and should be considered obsolete at this point.
A single E21A emitter already has a lot of tint shift to deal with; now we are adding 8 more of them, terribly off-center, and in the same optic.
Has anyone here ever measured CD/Beam distance using one of the 60mm Carclo Catadioptrics?
Here are some measurements, but with unusual or unnamed emitters which makes comparison difficult; Eye of Sekhmet - Exotic Hardwood and Copper - STO Flashlights - #19 by djburkes
Assuming that the throwiest option uses dedomed XP-G2, that’s 260klux (1020m) per 60mm diameter, which is 17 ANSI m per mm diameter. In comparison, a C8 with a 40mm reflector does around 800m with the same emitter, which comes out to 20 ANSI m per mm. Pretty efficient, pretty much as efficient as a reflector up to measurement error.
Thank you that is very helpful.
Have you ever seen a Catadioptric available in any smaller sizes, like 50mm etc? I cant seem to turn anything up.
I haven’t seen any smaller ones, and searches also returned nothing. I guess the main appeal of catadioptric optics is the flatness relative to diameter; below some threshold this advantage begins to disappear as conventional TIRs or convex/Fresnel lenses are sufficiently compact.
Are you thinking about a project that involves a catadioptric?
Reading this thread is making me suspicious of the performance claims from STO. The claim of 294kcd out of a reflector C8 is highly suspect–I haven’t seen even modern-day Osram and Luminus emitters measure this much candela in a C8.
It is also possible that the Osram Black Flat HWQP was used in the catadioptric light, but it’s hard to tell as they refuse to specify the emitter used.
In either case, the catadioptric optic may have much worse efficiency than previously estimated.
Eh it’s been long enough and I don’t make flashlights anymore so I’ll just tell you how I made that C8. It was a binned, hard driven OBF and of course focus spacers carefully adjusted. Then that extra little boost came from a tiny plano convex lens carefully glued to the center of the C8 lens with a diameter about that of the dead space in the bottom of the reflector where the LED sits and a focal length of course equal to the depth of C8 reflector. Probably works best with an flat LED that has an emission profile with lot of light directly outward but I went straight to the OBF for obvious reasons.
I don’t think that’s cheating to call it a reflector C8, the reflector is doing most of the work, that’s also why I never show the front. How did I find such a super specific lens? A lot of searching and luck, lol.
From tuning the catadoptric I recall it working pretty well, but you need to have the center alignment and spacing just right so really needs custom centering spacers, etc, etc. Oh and the Eye of Sekhmet had a variety of emitters depending on the version, the binned OFBs being one of them, no dedomed XP-G2s though, too unreliable, easier to bin OFBs from good bin OFBs.
Thank you so much for reaching out and explaining how you made that C8! I really appreciate the clarification, and apologize for being accusatory. Using a convex lens to hide the central truncation in the parabolic reflector is a very clever idea–I recall seeing this idea in many TIRs, including Carclo triples/quads. And thank you for confirming that the emitter in the Eye of Sekhmet is indeed an OBF.
Just a quick numerical sanity check: this test puts the Black Flat at 246 cd/mm^2 measured directly in front, so with a 40mm diameter optic one should expect a lossless intensity of 246(pi)(20)^2 = 309133cd. Djozz measured a typical reflector to be around 10% lossy; assuming negligible loss from an AR lens, the resulting intensity is 309133*0.9 = 278219 cd, which is close enough to the claimed 294kcd to be within the range of measurement error or binning variation. Ok, I can believe this measurement, the central convex lens (and the manual binning of the LED and optimization of focus) really makes a difference!
BTW, I noticed that the product pages for your lights were down only very recently–I remember checking out the Lance of Ra a few weeks ago and the page was still up. I am a huge fan of that light (despite not owning one) and keep referring back to it for my ongoing Fresnel thrower project. Sadly the Wayback Machine does not seem to have a copy of your pages.
Also great to see you back at BLF, even if briefly, the site indicates you last posted 9 years ago!
@Lumencraft-Matt Regarding the catadioptric: if I understood correctly, the above discussion indicates that both the C8 and Eye of Sekhmet are using the OFB, but the C8 out-throws the latter despite having a smaller optic diameter.
Without the central convex lens the C8 would be down from 294kcd to around 267kcd (assuming 12mm diameter truncation of the reflector), matching the Eye of Sekhmet, but with a much smaller optical area, by a factor of (40/60)^2 = 44%. So the catadioptric seems only half as efficient as a reflector for generating intensity.
My interest in the catadioptric is partly to do with how flat it lays, but I would also like to generate a beam that resembles an aspherical lens at peak focus (IE projects the die image), But I really only have about 50mm to work with.
I could make an “expansion” for the light to hold a 60mm optic, but that’s not ideal.
I ordered some of the 50mm 3 degree convoy optics to play around with, but I sort of doubt this lens is going to give me what want either. People say it has artifacts, and IDK if crystal sharp focus is going to possible or not.
