FW3A, a TLF/BLF EDC flashlight - SST-20 available, coupon codes public

So if someone wants to know - I have contacted Lumintop via AliExpress and they send me this glass for 6$

I bought my FW3A in November 2019 with SST20s and I EDCed it for a while but, believe or not, I put it away. I read something in a post that rekindled my interest and even ran a few tests. I was messing around with it in my 'box', after calibrating it, measured turbo, repeat turbo and max ramp. It looked like this and I'm wondering what causes the pattern circled in red... specifically the half-way point of the peak.

Thank you!

Thermal regulation.

Updated the top post:

Warning to potential buyers (2021)

The FW3A design has been changed many times since launch, with quite a few downgrades. It is not the same light any more. The 2021 model appears to be downgraded so much that it is no longer compatible with the upstream FW3A firmware. It’s missing an entire power channel, as shown in the pic below.

Additionally, the FW1A has been downgraded in at least two major ways — the reflector no longer focuses correctly, and the driver no longer has the primary regulated power channel.

These are only a few of the many downgrades though. Almost every time a new batch was produced, people reported surprising new changes, and most were not for the better. Several vendors have dropped the entire product line because it had too many problems and they had to issue too many refunds.

So, be warned: If you buy a FW3A, FW1A, or any of its other derivatives, you probably won’t get the light described in this thread.

Left: Original FW3A driver. Right: Downgraded version from 2020-12.

Total coincidence, but I just noticed this update was comment #13666, a combination of two of the most famous “unlucky” numbers — 13 and 666.

A great flashlight has apparently been ruined. A great flashlight that was not created by the manufacturer, who should have felt very fortunate. I find it disgusting.

Wow, I wonder what the newer driver design is. FET+1? Different MCU? Different firmware?

This is truly sad. You’d think that makers such as Lumintop would actually improve something created by a small project initiative. They took something great, watered it down, and then made it unreliable. I was wondering why we’d be seeing complaints from people when early release lights would behave just fine. Now… we know.

I had a feeling something was up when I saw significant changes going on with the BLF LT1. Notably, in the version produced by Sofirn. The LT1 v2 showed a number of board changes. If the FW3A is now neutered to the point where upstream software won’t work properly, what of the LT1?

Is the original design public? I wonder if someone else can start manufacturing it. There are also tail switched Anduril alternatives like the Noctilux KR1, though it is bigger than the FW3A.

Oh well, I’m looking forward to the SP10 Pro and hope it doesn’t get messed up like this.


Thanks for pointing this out.

I got an older FW1A (more than a year ago) and another FW1A a few months ago. Both are the SST20-4000k version.

They both have the driver as shown on the left side of your photo (not the "v1.1 2020").

Visually, the reflectors look the same, but I tried to compare beamshots, and noticed that the older reflector was more focused, while the new reflector has a bigger hotspot.

left side: Lumintop FW1A (SST20 4000k, older batch)

right side: Lumintop FW1A (SST20 4000k, fairly recent batch)

both of them use the same driver (as in the left side of the image posted by FW3A Team)

If I understand correctly, the new driver is a FET+1 and I think it uses the fw3a-219 firmware. At least, that’s what I’ve heard from people who got one recently.

What that means is:

  • No more regulated power between 0.35 A and 2.8 A.
  • Ramp levels 66 to 130 are DD FET, not regulated.
  • More heat, less efficiency.
  • Ramp levels 131 to 150 are pointless, since they’re all full-power. It tries to ramp the third channel there, but that channel doesn’t exist.
  • Thermal regulation won’t begin to reduce power until the light is significantly hotter, since the first 20 steps of reduction (150 to 130) do nothing.

I’m guessing the driver was changed to reduce costs again (cha bu duo), but someone noticed turbo didn’t work without the third channel. Normally, channels 1 and 2 (regulated) are turned off at the turbo level, because a DD FET (channel 3) is brighter by itself. But after moving the FET to channel 2, no more turbo. So they switched to the 219 firmware instead, since it keeps channels 1 and 2 on at turbo. Perhaps they don’t expect anyone to notice if there’s no difference between the ramp ceiling (130) and turbo (150).

The original project got delayed for like a year because Lumintop couldn’t meet the specified quality standards… but then it went to production anyway, and there were frequent quality issues. Vendors reported a really high rate of returned items, and Illumn had to manually adjust and test each light before shipping. So it didn’t get carried for long.

Over time though, parts kept getting changed, usually to cheaper versions. Like, for a while, the Carclo optic and lens got replaced by a one-piece plastic imitation which was a huge downgrade. There was a lot of backlash from that though, so the lens was brought back and the optic was redesigned to be a bit closer to the original.

Another frequent issue was that the driver cavity frequently wasn’t big enough for the driver, so a bunch of lights were made with the driver sitting on top of where it was meant to go, instead of inside its compartment. This made the connection flaky, and made batteries not fit right, and was one of several sources of unreliability.

The springs got downgraded at some point, using a type which had some sort of coating which flaked off easily, and got flecks of metal all over the inside of the light. This caused a bunch of electrical shorts, and more unreliability.

The button got downgraded from its original ~800g pressure to only ~300g, and sometimes even less. This made it hard not to activate by accident, and people had to carry the light in lockout mode.

The entire tail end got redesigned several times, and I don’t know if it was ever fixed to be reliable. It’s famously finicky.

A bunch of spinoffs were made, and some were embarrassingly lazy. Like, the 21700 version seemed like it was simply scaled up without adjusting proportions… so the battery compartment ended up so long that it didn’t fit the cell it was meant to use. Because 18650 scaled up to 21mm width… is 21x76mm, not 21x70mm. So a common 21700 cell needed a 6mm spacer in order to make contact. The clip wasn’t adjusted either, so it was too short. The optic was simply reused, instead of scaling it up for better throw… and it had a bunch of empty space around it. Etc.

And now half the driver hardware is missing. I guess it’s more of a “FW2A” now.

The problematic tube design was never corrected. Mention it and a handful of people will instantly tell you that you’re too stupid to fix it yourself.

With so many fanboys, why should Lumintop make this light better?

Wow. That’s more than a change, that’s just a total nerfing of the light. What a shame. I guess I’m glad I got in on the first couple batches…mine has been perfect as has the one I gifted. I’ve wondered about all the problems and the switch issues/mods. I didn’t realize so much had changed beyond that initial optic swap fiasco.

Do you suppose this driver thing now is just cost cutting or could it be a temporary thing with chip shortages?

wow thats just sad… nerfing the light and quality just to maximy profits… thast now how u get happy customers lumintop… :person_facepalming:

im so happy i have mine from the group buy but it sucks for newer buyers that want a blf designed light only destroyed now… sad

This sucks.

Sorry to hear about Lumintop cheapening what was a classic light.

Thanks for heads up.
Lucky me, I got FW3A from first batch, clean, without any markings :slight_smile:

What supplier was this red driver version from? I was considering buying one more, so I guess I oughta do it soon.

The downgraded model in the picture was purchased through the Lumintop official aliexpress store.

Other people have reported that they received older versions recently though, like the old driver with Anduril 1 and all the chips. So there may still be some old stock floating around.

So essentially, the way Lumintop has handled this, there’s no clean rotation of stock. I wonder how many Anduril 1 lights are still sitting around that will end up sold as Anduril 2. I wish there was some exterior visual cue to help with this… but at least there’s the driver board visible after removal of the head from the body tube.