TLF/BLF/Lumintop FW3A review (18650, 3x XP-L HI cool white)

The FW3A on turbo with a 30Q heats up to about 50°C when the thermal limit is set to 55°C. It stabilizes at around 46-48°C. This is a good setting for not burning your hand. Also loving how quickly the heat transfers to the battery tube from the head.

In case we do end up with the slightly older version of the firmware (not that it sounds like a big deal), the FW3A has accupuncture style flashing holes, right? At least, I think that’s what I heard but I don’t have time to check at the moment. I’ll search some posts later when I get time, assuming nobody has already answered by then. Reflashing should be somewhat easy if this is the case though.

No flashing vias on the driver backside from what I can see.

Great review! :slight_smile:
Thank you so much for your work.

One little thing

Do you mean non-magnetic ?

Edit: I should read first the whole thread, it’s already mentioned

Kiitos suuresta työstä, maukka! :+1: :slight_smile:

The Emisar D4S is flashable like that.

Thanks for your review maukka, looking forward to your beamshots/test data with the correct optic (10511) installed.

Happy to report that the beam is very pretty with the 10511 optics. No discoloration in the middle of the beam. Just some radial artifacts at the edge of the beam, which are not even visible on a white wall unless you rotate the light.

Sounds about right. It uses a window of 10 C as its target zone. If it can get the temperature anywhere in that window, it stops trying to adjust the output. So with a limit of 55 C, the window would be from 45 C to 55 C.

If you happen to have some other optics like 10507 or 10509 it would be nice to have some pictures, all from the same distance, to see what they look like. Assuming your not too busy.

That’s more like what I was expecting, thanks for checking and getting your findings up so quickly :+1:

A quick note about lumen numbers:

Maukka uses proper ANSI lumens, but a lot of the flashlight industry uses a different scale which has more impressive-looking numbers. I call this second one the “cpfselfbuilt scale” because he was the one who really perfected it. He did this by measuring a lot of lights, trusting that the manufacturer’s values were correct, and fitting a curve to the average. So it does a pretty good job representing a lot of brands which were popular at the time, even if it’s not really accurate. Anyway, the difference between the two is roughly 30% or 40%, so 100 ANSI lumens would be about 130 or 140 cpfselfbuilt lumens.

My measurements, so far, have used a super low-tech “integrating milk carton” which I calibrated to match cpfselfbuilt’s numbers as closely as possible. One of these days I’ll buy some reference lights from maukka so I can get more accurate numbers, and perhaps build a proper integrating sphere… but that hasn’t happened yet.

When I measured my FW3A production sample, I got the following results:

  • 1x7135 level: 155.8 lm
  • 8x7135 level: 1163.8 lm

… and maukka’s numbers for these same levels are:

  • 1x7135: 119 lm
  • 8x7135: 888 lm

So in this case, the difference is 31%. That is, assuming the two lights actually produce the same amount of light, which is questionable because the emitters are different tints with unknown bins.

Maukka also measured 2800 lm at 0s on turbo. Using the same 31% correction factor, this works out to about 3666 lm on the scale I’ve been using. That number is probably more directly comparable to the ~4300 lm I got with an Emisar D4.

All the FET / turbo stuff is going to be incredibly approximate though, not precise at all. The actual values are highly dependent on analog factors like the battery condition, the cleanliness of the contacts, etc. But mostly I wanted to let people know that 2800 ANSI lumens is not a disappointing result. It’ll probably look almost as bright as a D4. Unless they’re side-by-side, it would be hard to tell the difference.

Re: Flashing vias - Ah okay, I must have been thinking of a different flashlight then. I know the D4S has them, and I think there was a rather premium keychain light project that was going to have them too. That might have been what I was thinking of.

The fifty shades of grey anodizing is looking really nice. It’s silvery reflections can fool you depending on the lighting condition. Looking at the short angled area of the body near the clip… In one pic it’s the brightest area of the body, but in another pic this area of is turned away from the camera, and it’s so dark that at first glance I thought “Why is there an extra O-ring there?” It’s even better because something that could have been a major drawback of the light turns out to be a superior point.
It’s not as good as the naked prototype, but it’s as good as it possibly could be.

For comparison, here’s some lights so you can get an idea of the shade of grey of the FW3A

Beamshot comparison with different optics. The difference between 10508 and 10509 in the photos is very small, but measurably significant. It is possible though, that the optic installed in the light was 10509, since I don’t have a 10508 to compare it to.

GIF

Individual beamshots

Nice. Thanks

Very nice review for a very nice flashlight!

Question regarding updating the firmware, since there are no flashing vias, how does one update the firmware if that’s possible?

You’ll have to pull the driver out and attach the clip to the MCU (like normal).

Is that glued or soldered? Any idea how difficult it might be to remove and to reassemble back?