Quick guide to Carclo optic selection

Flashlight Modding and DIY Parts
1

For those who are confused by Carclo’s weird numbering of optics and misleading descriptions, this is to help you decide on the optic you want! [This is a crosspost for the purpose of easy access.]

Different Carclo optics produce different beam profiles due to different TIR geometry and different surface treatment. The triple and quad optics use the same individual TIRs, and matching TIRs produce the same beam profile. For those who are curious, the triple/quad correspondence goes:

  • 10507 (triple) is 10621 (quad)
  • 10508 (triple) is 10623 (quad)
  • 10509 (triple) is 10624 (quad)
  • 10511 (triple) has the same geometry as 10622 (quad), but with light frosting.

If polished to a smooth finish, the 10508/10511/10622/10623 have the same underlying optics, just with different levels of surface frosting.

So which one should you choose? That depends on what beam characteristics you want.

  • For throw in domed emitters the 10507/10621 generally make the best beams, with a well-defined round hotspot (no corona) and virtually no tint shift within the hotspot. The Acebeam E70 mini pairs domed 519As with the 10507.
  • For throw in domeless emitters the 10511/10622 generally make the best beams, with a well-focused hotspot and generous corona, without too much starburst artifact.
  • For flood, 10508/10623 are good for making a smooth cutoff without too much tint shift.
  • For rosiest hotspot tint, go with 10507/10621, everything else has varying levels of tint shift that makes the hotspot a bit yellow.
  • For an elliptical (elongated) beam, the 10510 triple is an option. This might be useful for bike lights or whenever a wide strip of area lighting is needed.
  • Artifacts: if the emitter has a phosphor pour that extends past the die (such as E21A and XPG3), or otherwise severe artifacting or tint non-uniformity, no clear optic can smooth it out. A frosted optic must be used to smooth out the artifacts, and throw must be forfeited. Thank you to @cannga for pointing this out.

I do not generally endorse the 10509/10624, the tint shift makes the beam from most domed LEDs unusable. Use only if the LED gives a blue hotspot and yellow spill in a reflector (i.e., has tint shift in the opposite direction as regular LEDs), such as Nichia Optisolis.

If you live in the US, the best place to get these is LEDSupply, which sells triples for $1.5 and quads $2.38. With a $3.99 shipping fee, you might as well sample all of them!

15 Thanks
2

Yeah, I got most of my TIR lenses (17/20/26mm, not necessarily triples/quads) from LEDSupply, and all those various pages for each were having me go in circles.

3

Wait. Are you sure the 10511 and 10508 have the same underlying optic? I thought it was 10507 and 10511?

4

The 10507 has physical divots(or whatever they are called) in the top surface, none of the frosted ones have that:

Screenshot_20240405_234151_Chrome
Screenshot_20240405_234151_Chrome1016×1186 136 KB

1 Thank
5

Yes! The 10507 has a construction that is entirely different from the other optics in the triple series. The central convex lens in each optic is much larger, which is what makes the hotspot tint more rosy.

I find Carclo’s nomenclature misleading: if the 10507 is called narrow clear and the 10511 narrow frosted, one would expect them to be the same optic just with different surface treatment, and that is what I initially assumed, but that is not the case! What’s worse is that the official blueprint fails to reflect this difference in dimension, instead showing the smaller central lens of the other optics. Hence this post.

2 Thanks
6

That is very interesting information! I didn’t know that at all. I think I’ll keep my 10508 and polish it a bit too get the 10511 effect then!

7

Maukka’s FW3A review has a ton of great info and measurements about Carclo’s triple optics.

I find that 10511 is the best choice for most lights, and 10622 is the best for most quads. Usually the optics with a flat top make the nicest-looking beams, while the optics with parts scooped out (10507 and 10621) make the ugliest beams with the most artifacts and tint problems.

Meanwhile, the extra-frosted flood optics make the smoothest beams but the intensity is so much lower that it requires doubling or even tripling the power (and thus cutting the runtime in half or a third) to reach the same apparent brightness.

For more throw without sacrificing beam quality, the 10511 can be polished pretty easily by rubbing it against a pants leg or something, to wear down the frosted surface. This gets the throw up to about the same level as 10507, but with a much nicer-looking beam. Or just use it as-is, because the throw is already pretty decent.

From his review…

5 Thanks
8

I was generally aware of the beam characteristics of the Carclo triple optics. I didn’t know that the 10511 and 10508 were the same optic with different frosting levels (I had assumed it was the 10511 and 10507).

I also didn’t know about the characteristics of the quad optics, so that is good information too!

9

This is very good info, and I certainly agree with the analysis for domeless emitters. However, care must be taken to acknowledge that the analysis does not generalize to domed emitters like the 519A.

Domeless emitters barely have any angular tint shift issue to deal with, so a good beam profile can be prioritized over tint consistency, making 10511 preferable.

For domed emitters, however, the issue is the opposite: the more rounded light-emitting surface minimizes beam profile artifacts, but the dome introduces a nontrivial amount of angular tint shift that needs to be dealt with by the 10507.

These Carclo optics are a combination of a traditional TIR (which functions like a reflector minus the spill) surrounding a core that is a convex lens. With the 10511, the central lens is quite small, so the beam takes the form of a large “corona” from the convex lens, which is rosy, plus a central hotspot from the surrounding TIR, which is yellow-green, hence the tint shift. With the 10507, the central convex lens is larger, which makes the rosy “corona” shrink in size and increase in intensity, simultaneously making the hotspot rosier and making the rosy corona disappear. The shrunken rosy corona is about the same size as the yellow-green hotspot from the outer TIR ring, which results in a single, sharp hotspot that is uniform in tint.

10

For domed emitters, I still don’t like the scooped-out style optics like 10507.

For simplicity, I’ll use Carclo’s more recent terminology of “top lens” vs “plain lens” to describe the scooped out style (10507 + 10621) vs flat style (10511 + 10622). Not sure if that’s the right word for it, but I’ve seen them use “top lens” on most of their scooped models.

I have lights using a variety of different LEDs in a variety of different optic styles, and there is only one case where I prefer a top lens model over a plain lens. That’s when I use an Osram W1 – because the point of that LED is to maximize throw at any cost.

To confirm, just a moment ago, I checked a top lens domed 519A against a plain lens domed 519A, and the difference was pretty dramatic. The top lens made a beam I find ugly and impractical, basically just a weird-colored hotspot with colorful rings around it… almost like using a zoomie. Meanwhile, the plain lens made a balanced spot+corona beam with good color consistency.

I’ve tried a bunch of others, and for me, plain lens optics win every time – with the sole exception of an Osram W1. Because for that LED, I don’t care about the beam quality, I only care about getting the center of the hotspot as bright as possible.

So I’ve removed top-lens optics from as many lights as I can, replacing them with plain lens optics instead, and basically never use the ones I haven’t been able to fix yet.

11

Which triple has the overall best throw with any LED?

12

10507 gives the most candelas.

I hate how it makes the beam look, but if throw is the only factor, it’s the top choice.

1 Thank
13

It’s surprising that we have such different experiences with these optics! I recognize that cameras might be poor at picking up this sort of detail, but do you happen to have beamshots showing the weird-colored hotspot with rings?

It looks like opinions vary on the zoomie-like beam that the 10507 produces, some folks really seem to like the Acebeam E70 mini beam profile, which uses 519A with 10507.

14

As you say, it’s hard to show on camera… because colors are hard to show accurately and the dynamic range is way too high for most camera sensors. But here’s a shot of unmodded (domed) 519As in top-lens Carclo optics.

The beam has at least 4 different sections, each with a different brightness, tint, and color temperature. The hotspot is bright and off-white, with a relatively sharp drop-off. Then there’s a dark ring with a very different color and temperature. Around that is a lighter ring with more of a neutral color and temperature. Then around that is a wide, dark spill area.

The tint and CCT shifts throughout the beam make it so none of the beam looks neutral. In much the same way that two different lights next to each other makes one or the other look tinted, the top-lens beam makes itself look tinted.

When I put the same domed 519A LEDs side by side with top-lens vs plain-lens optics, the plain-lens beam is neutral and smooth, while the top-lens beam is visibly more green.

These are the same LEDs… just different optics. Top lens style on the left, plain lens style on the right. It changes the color quite a bit.

The left has a harder edge around the hotspot, with a somewhat higher cd/lm beam. It’s hard to tell since the image is blown out and clipping, but the left hotspot is brighter and off-white… while the right hotspot is slightly dimmer and neutral.

Normally they wouldn’t be so close to what they’re shining at, but to get both in a single picture, I had to get pretty close.

Here’s another domed LED (XP-G2) in a top-lens 10507 optic:

This was a while back, but on the right is a larger LED (LH351D) in a 10507 optic. It had a particularly zoomie-like beam, with a big flat-looking circle and a sharp colorful edge. On the left is a 219B in plain optics.

Here’s an Osram W1 in a top-lens throw optic. It may not be pretty, but it sure does throw well:

That’s the only light I actually want a top-lens style optic in. Because on that particular light, I only care about throw.

1 Thank
15

Well. I did some polishing to a 10508 Optic. It likely increased throw and made the beam more useful, but I noticed a bit of tint shift in the very center. Pretty mild. That said, this is definitely not the same as a 10511. Beam angle is definitely different. Floody of course. The structure may be the same as a 10511, but it’s not the same optic.

16

Thank you for making the effort to send these photos! I agree about dynamic range and color, and also sometimes tint can change with intensity on crappier phone cameras.

I am extremely surprised by the amount of green I’m seeing, the difference is like a normal 519A versus SST40. I do not observe this with any of the 10507 beamshots with 519As, like with the Acebeam E70 mini. BTW what lights do you have here? In order for the beam to turn this green, a massive amount of red and possibly blue must have been lost (absorbed or scattered by the optic), but I could not think of a single mechanism by which this can happen.

Qualitatively I can imagine this happening because the flat optics have a very pink spill like a reflector, but these photos seem to indicate that the hotspot itself is greener–is that what you observe irl? I really wish I could check these lights out in-person, do you have guesses on what makes one of them so green? Or maybe the tint difference between hotspot/spill is confusing the sensor? It’s also confusing to me that the sensor seems to find the low CRI XP-G2 less green than a 519A; almost every camera has it the other way around.

I am personally ok with the tint shift and artifacting in the spill–the intensity is low enough in comparison to the hotspot that I don’t notice outdoors and can choose to not notice indoors; for me the tint consistency in the hotspot outweighs all. But I still don’t understand why your optic is messing up the tint to such an extent.

17

I tested my 10508 and 10511 by using clear tape to smooth out the frosting and make it perfectly clear (as the glue goes between the tiny frosting beads), and got essentially identical beams. Let me test again and see if I can get some decent photos (might be hard as my current phone has the worst camera I’ve seen since 2014).

18

Really?! Maybe I have a 10509. I’ll need to check again.

19

Just checked, the smoothed 10508/10511 did produce the same beam: a distinct central hotspot that is noticeably yellower than the corona, which has a distinctly triangular shape when rotated and is peach-colored with some different-colored phosphor patches/sprinkles visible. BTW if it’s an 10509, the holes where the LED domes go are smaller/shallower compared to 10508/10511.

Tried 10 minutes to get some decent photos and simply gave up on decent. Even with the same optic and same beam, every photo shows a different beam shape, different tint, and different tint distribution; on top of this, the tint varies wildly with intensity. Sigh, I will upload them as an example of why beamshot photos often can’t be trusted…

20

imgonline-com-ua-twotoone-qumNGOBWsd
imgonline-com-ua-twotoone-qumNGOBWsd1920×1280 60.2 KB

Left: smoothed 10511; right: smoothed 10508, both 519A 3500K. I chose these two beamshots because they manage to show the central hotspot.

The tint within each photo is completely messed up; the camera artificially perceives steep intensity gradients where there are none, and messes with tint along these boundaries. To illustrate, here’s the same photo with increased saturation; look at how brown the boundaries are!

processed
processed1920×1280 71.3 KB