"The LOUDENER" 4.66 km ANSI, 230mm fresnel mod for Convoy C8+/M21A

[Haedesz]

I’m excited to present my ridiculously pocketable thrower project!

It’s a Convoy M21A (or C8+) host with the SFT-42R and standard 3V10A driver.
The cone has been 3D printed out of PETG, together with a printed retaining ring and the lens is a 15$ 230mm PMMA Fresnel.
The cone slides on three very secure rubber O-rings, which technically makes it a zoomie!

# First conservative readings are around 4.2mcd ~ 4100m ANSI throw (2.55 freedom miles).
Update 12 October: I did another reading after fiddling with it some more, we are now up to

5,431,940 cd ≈ 5.43 Mcd or 4660m ANSI throw!

The beam itself is very unconventionally looking. It converges and becomes smaller, instead of “opening up” like a traditional beam. Very death-star like!

but that’s just by eyeballing the focal point, no real maths went it to it (yet).
Personally I have never seen such a tightly collimated (LED) beam IRL in my flashlight history.
The pictures are taken in the evening, it wasn’t even fully dark out yet! But I couldn’t wait to share .
Lighting up the clouds with the shots.

Next I’m going to print a lens head for the 3x21D (F150R with 25A driver)

Reddit thread

Printables files and instructions

Specs (approx):

3000-ish lumens (SFT-42R / 3V10A buck driver) 
~4.2 million candela equal to 4100m ANSI throw (2.55miles)
Convoy C8+ or M21A modification 

BOM:

Convoy C8+ or M21A (preferably with a round emitter)
Printed cone and press-fit retaining ring (preferably PETG)
Fresnel lens Diameter = 230mm, Focal length = 190mm
3 Rubber O-rings (I used 32mm ID, 3.55mm thickness)
Matte black paint 

Instructions:

Print the cone and retaining ring
Paint the inside of the cone matte black (PETG is inherently shiny)
Mount D= 230mm F= 190mm fresnel lens inside cone and press fit retaining ring
Carefully remove the bezel AND reflector of the C8+/M21A
Mount 3 O-rings inside the 3 "heatsink" notches of the C8+/M21A's head
Slide cone over the O-rings.
Adjust beam convergence by sliding the cone up/down (shouldn't take much, best done outside)

21.10.2026 - LHP73B version of the M21A:

It’s definitely a lot “louder”. More brightness, bigger (and square) hotspot but less throw. Around 3.5mcd which equates to 3,742m FL1 ANSI throw.

Also got nice overseas mail from @contactcr!
Some ancient tech which isn’t seen a lot anymore:
Original Marinebeam Wavien Collar (2018 group buy).

This will be used in the 3X21A build which also arrived today, the funnel is printing right now!

22.10.2026: I finished the 3X21A prototype!
Preliminary results (without wavien) sit at:

6.37 million Candela ~ 5050 meters FL1 ANSI throw. With an estimated max of 1800 lumens.

This breaks the magical 5km barrier where the Acebeam W50. 2 multi thousand $ LEP sits at.

This concludes part two of the project, the last update will include the Wavien collar, if I ever get that to work consistently!

https://imgur.com/gallery/storm-clouds-sQ2OkRb

Special thanks to @qreciprocity42 for theorycrafting and @Simon_Mao for making such awesome hosts!

[raccoon]

Nice!
An alternate name could be “The Embiggener” (from The Simpsons).

[Haedesz]

@raccoon

Nice!
An alternate name could be “The Embiggener” (from The Simpsons).

Another Simpsons fan

[QReciprocity42 (efficacy ≠ efficiency)]

Fantastic mod and writeup!

@Haedesz

lens is a 15$ 230mm PMMA Fresnel.

Where did you buy this? I got a 200mm one for $16 on Ebay, and would love to learn of a cheaper source.

@Haedesz

but that’s just by eyeballing the focal point, no real maths went it to it (yet).

The 4.2Mcd does seem a conservative estimate. Assuming 200cd/mm^2 from the emitter and a 80% loss in luminance through the optical system, the 230mm diameter should give 200*pi(230/2)^2 = 8.3Mcd in principle. But perhaps factors such as chromatic aberration contributed a lot to the loss.

[EDIT: botched calculation, see correction here. The conclusion of underestimation still stands.]

When you build this for the 3x21D with a larger emitter, chromatic aberration and other optical imprecision issues would be less significant, and I would expect >5Mcd.

Did you end up removing some of the material from the M21A’s head to widen the emission cone, so that it covers the entire Fresnel lens? Or was removing the bezel already enough?

@Haedesz

Personally I have never seen such a tightly collimated (LED) beam IRL in my flashlight history.

With a 2.32mm emitter diameter and 190mm focal length, the beam divergence angle is roughly 2.32/190 radians, or 0.70 degrees. Certainly tighter than pretty much every commercial LEP out there, but still slightly wider than a full moon.

@Haedesz

3000-ish lumens (SFT-42R / 3V10A buck driver)

Since the lens only collects part of the LED’s emission, the actual total output is sadly lower. Assuming a Lambertian emitter, a D230 F190 lens only collects ~26.8% of the output (without accounting for reflection+scattering losses), so the total lumens cannot exceed 800-ish lm.

[Haedesz]

@QReciprocity42

Fantastic mod and writeup!
Where did you buy this? I got a 200mm one for $16 on Ebay, and would love to learn of a cheaper source.

I am not sure if I can freely post Aliexpress links here but this is the lens I bought:
Aliexpress link
It is not perfect (marring) and came rather dusty but in the end, it still makes me smile
Packaging was good though: paper film protection > cardboard > wrapped in atleast 2 meters of bubble wrap > padded enveloppe.

@QReciprocity42

The 4.2Mcd does seem a conservative estimate. Assuming 200cd/mm^2 from the emitter and a 80% loss in luminance through the optical system, the 230mm diameter should give 200*pi(230/2)^2 = 8.3Mcd in principle. But perhaps factors such as chromatic aberration contributed a lot to the loss.

I do not have any “real” testing equipment, I just used the luxmeter on my phone and calculated the cd from there. It guesstimated the 3X21D candela within 10% of spec before - So I think its correct.
And like I said in the post, the focussing was just done by eye (what looks right), definitely some margin for user error.

@QReciprocity42

When you build this for the 3x21D with a larger emitter, chromatic aberration and other optical imprecision issues would be less significant, and I would expect >5Mcd.
Did you end up removing some of the material from the M21A’s head to widen the emission cone, so that it covers the entire Fresnel lens? Or was removing the bezel already enough?

The moment the lens arrived, I just mounted it- I needed to SEE . So its just the standard aperture of the existing head (this will probably impede some of the light reaching the lens)
But I’m so impressed with this build, that I’m not going to alter the head for now - I’m just going to convert it fully and correctly for the 3X21D and try to finetune it there.

@QReciprocity42

With a 2.32mm emitter diameter and 190mm focal length, the beam divergence angle is roughly 2.32/190 radians, or 0.70 degrees. Certainly tighter than pretty much every commercial LEP out there, but still slightly wider than a full moon.

I was very surprised to see such narrow beam IRL. That’s why I’m confident to at least try the larger F150R for the next build (same lens).

@QReciprocity42

Since the lens only collects part of the LED’s emission, the actual total output is sadly lower. Assuming a Lambertian emitter, a D230 F190 lens only collects ~26.8% of the output (without accounting for reflection+scattering losses), so the total lumens cannot exceed 800-ish lm.

Yes, I sheepishly reffered to the “standard” output of this emitter in most builds, thanks for correcting me!

====

BTW, The camera does NOT pick up the beautiful colors that the beam goes through when collimating. I’ve seen blue, purple, magenta and pink consecutive edges around the hotspot.

[QReciprocity42 (efficacy ≠ efficiency)]

@Haedesz

I am not sure if I can freely post Aliexpress links here but this is the lens I bought:
Aliexpress link
It is not perfect (marring) and came rather dusty but in the end, it still makes me smile
Packaging was good though: paper film protection > cardboard > wrapped in atleast 2 meters of bubble wrap > padded enveloppe.

Thank you very much for the link! I post them occasionally when relevant, and haven’t run into issues. You must have gotten it on discount!

I got the exact same packaging with my Ebay lens. I bet they all come from the same supplier. The dust is unfortunate, especially if it gets stuck in the grooves.

@Haedesz

I do not have any “real” testing equipment, I just used the luxmeter on my phone and calculated the cd from there. It guesstimated the 3X21D candela within 10% of spec before - So I think its correct.
And like I said in the post, the focussing was just done by eye (what looks right), definitely some margin for user error.

I see! I also measure with my phone and calibrate to 1Lumen measurements using known lights. Throwers are usually underestimated, sometimes by a very significant amount, when measured at close range.

Best of luck with the 3x21D conversion! It’s gonna be fantastic.

@Haedesz

BTW, The camera does NOT pick up the beautiful colors that the beam goes through when collimating. I’ve seen blue, purple, magenta and pink consecutive edges around the hotspot.

That’s the chromatic aberration! Yeah, the camera doesn’t have enough dynamic range to pick it up next to the hotspot. Unless the whole hotspot is pure chromatic aberration!

[Ledhead]

I thought of this when I saw this thread: It makes me wanna SHOUT! A little bit Loudener now! @2:30
That’s crazy fun!

[Henk4U2]

Amazing! Next question is of course: where to source enough wool to knit a holster.

[Dc38 (The Glutton)]

Spin that yarn about the far-reach of The Loudener, the holster will weave itself!

[lndger]

so cool. Very interested in the 3x21d version. Will you be reusing that same 230mm lens?

[Haedesz]

@lndger

Will you be reusing that same 230mm lens?

Thank you!
Yes, I will be reusing the same lens for now. I can only print max. ~260mm in one piece, so for 300mm lenses I would have to print the head out of multiple parts which is not worth the hassle right now.

I think for V2, it would benefit more to increase efficiency, output, light collection (and the quality of the lens, but that would be very expensive and too heavy). That’s why I’m going with the ““small”” LES high intensity F150R driven at >20A, and am hoping to get my hands on a wavien collar to recycle some of the lost side emission.

[lndger]

nice. I’m looking forward to it. I might try to emulate what you’re doing, but with a 300+mm Fresnel since my H2S has the build capacity. strong suspicion I’ll bungle it somehow, though, ha.

[Haedesz]

@lndger

I might try to emulate what you’re doing, but with a 300+mm Fresnel

Would love to see it , let me know if you need any help!

[Shirtgrinder (Matt)]

Biggus Torchus in a Michael Palin- type voice.

[raccoon]

Nice!
I had to look up Michael Palin.
He was in Monty Python, as expected.

[Shirtgrinder (Matt)]

From the movie Life of Brian.

[raccoon]

Yep, I’ve seen the movie at least once.
I also have the soundtrack and listen to it on a regular basis.

[Shirtgrinder (Matt)]

It’s a classic.

[BobbyMK]

Nice one, this reminded me of my frankenlight, the ugliest flashlight ever created, and it was by me back in 2019. Mine used much smaller aspheric glass lens, 70mm in diameter, the base flashlight for this ugly contraption was Haikelite HT08, the irony in all this is that the stock HT08 came with fresnel lens from factory.

It gave me idea, i still have the glass lens used in this one, might adapt it on some other flashlight.

[Haedesz]

@BobbyMK

It gave me idea, i still have the glass lens used in this one, might adapt it on some other flashlight.

That would work with the same type of adapter (cone) I used, but it would make it very front heavy. PMMA lenses are just so cheap and more importantly, light!