What did you mod today?

Today I modded my just received FWAA by removing the driver side spring and replacing it with a disc contact. This allows me to run my Shockli button tops that wouldn’t work with the spring in place. Shocked how small the light is honestly. Beam is nice for indoor use, don’t know how much I’ll carry this light honestly, but we’ll see. Considering swapping emitters to 219b or a tint mix of SST20s or even giving the HL2X 90CRI emitters a try that are collecting dust…

Also in the package was a FW21 X1L. I’m surprised how much I like the host. Looks and feels great in person. Even the ano color is somewhat pleasing on this body (not so on the FWAA - it’s about to get stripped or baked). The optic generates a super clean beam that I really like, but throw is disappointing… only 115kcd. Not going to cut it for me. If I can’t get that up to near spec’d range, swaps will happen. My Convoy M2 does 102kcd and is more pocketable

But really. The UI on this thing is AWFUL! Like wtf year is it? Long press for off, single to advance poorly spaced modes… no, no and no. This will have to change. Already taking measurements to convert this into an ANGIE quad. Hoping to get a driver designed for this running Anduril.

Is it actually using a different driver? That would make zero sense. Ideally it has a tiny85 on it, you can just put Anduril back on.

Thank you! I’m definitely going to look at those optics. I love my FW’s and would love more with single emitters.

Btw, what did you use to space up the pcb?

Still not finished but it will be.

New Armytek Prime C2Pro , still 12V driver so XHP35 and Yaijamei optic will do

Olight Warrior Mini 2 emitter swapped with SFT40. End result is great. Although i’m high cri fan, this emitter looks realy nice under Olight optic.


In my case I didn’t space it up, I machined my own bezel that’s a few mm shorter than the factory FW1A. I also had to machine the lower part (last ‘fin’) of the reflector off to fit within the head cavity.

Taking a second mcpcb, torching it until the top layers separate (doesn’t always work), then filing and sanding it smooth is an option to create your own spacer. Or buying copper discs from Bopper on Etsy is another (easier) option.

Yes. It’s a totally different driver and mcu. No markings and it’s small with ten pins IIRC, don’t quote me though. I’ll be tearing it down again… I did swap a Convoy Boost HX in. Throw dropped to 85kcd. Might try a 3000k SST20 next to see how it compares to my LEDIL IRIS in my ET G25C2.

Driver OD is 25.4mm. Cavity is 20.4 mm, so it mandates a 2.5mm ground ring. I wish it used the same driver as the X9L, I missed this detail and am kinda annoyed about it.

Just use an actual green emitter instead like Osram. At least the green will be consistent that way… :person_facepalming:

:smiley: I can’t remember how bad they were. I know the 5790 when sliced basically down to the phosphor, the angular shift was fixed and the overall tint was very nice for a daylight cct. 4000k never got much attention, likely due to impossibly high duv. I might just need to refresh my memory.

Small with 10 pins sounds like what appeared on the new Convoy drivers, too. Some weird Sonix 8051 clone. The curiosity is killing me.

It does appear to be the same chip as Convoy. Good catch.

Not really much of a mod, but I have to share. I put a Zircon 803 filter in an SP36 with 4000K LH351Ds and got surprising results. I’ll just let the data speak for itself.


Does Ra = cri …?

Yes, CRI (or Ra) is the mean of R1-R8.

Wow, cri doesn’t even include the much talked about r9 … :stuck_out_tongue:

thanks! nice data capture

The zircon 803 has a 20% lumen penalty
the benefit is the DUV went negative… congrats on that

for reference, could you also post the same light without the -green filter?

what I like about the SST is the high R9… but I dont care for the green tint

the Lee filter keeps the high R9, and improves the tint… similar to the way slicing the dome of an LH351d does. But the LH351d is not a 90+ R9 CRI LED… and the green tint is as bad as the SST…

so I just ordered an SST light… though I may still mod to 219b, instead of using minus green filters.

That is why R9 is often mentioned separately.

R9-R12 are saturated colours and were all not part of CRI because CRI represents how natural colours look, not how saturated. They are important for good lighting though, and with leds the R9 (saturated red) is always the bottleneck so that is the one mentioned.

Okay that makes sense. I just thought it was discussed because it was always lacking…

With the Zircon 803 installed over the LH351Ds, R9 came in at 96 as you can see in the data above. That’s what I found most impressive since the LH351D typically comes in around R9 60-70.

I’ll retake measurements without the filter when I get a chance.

thanks for clarifying, I was confusing SST with LH351d
that IS a big jump in expected R9 for an LH351d,

great visuals, love your color graphs

CCT Duv CRI R9
C8+ 351D 4000K Without filter 4200K 0.0039 91 70
C8+ 351D 4000K With filter 4060K –0.0019 97 94
D80v2 351D 4000K Sliced 3800K 0.0012 93 77

Edit: added D80v2 with sliced LH351D