FREEME ✌ ASTROLUX FT03S SBT90.2 4500lm 1428m 26650/ 21700/ 18650 Flashlight - EXPIRED

No clue -
Replacing wires and anduril are doable enthusiast fixes -
Resoldering is fairly simple.
Reflashing anduril is possibly somewhat easy, We’d need a better suited version of anduril, perhaps the MF01S version. There are pogo pads, which is great, but they’re not in the same configuration as the emisar lights, theyre 6x1 as apposed to
00
0000
which is obnoxious by honestly not that bad. either make your own row, solder tiny wires to it, or do a combo, hit 4 with with pogo pins, solder two.

Reflector fixes without them producing a proper reflector is essentially impossible for us.

Speaking of the reflector, mine is as expected the “bad” reflector, but by doing the “look into the lens and see how much of the reflector shows the diode at a few feet away isn’t doing the ”center of reflector isn’t showing it” which is what I heard I should see. Anyone have a photo of what that should look like not with other lights, but with the FT03S? Maybe some of the “bad” reflectors aren’t bad, or the led height is slightly different?

I don’t have measuring equipment, and the only throwers I have to compare hotspot size to are a SST-20 3000k D1 and D1S, and SST-40 D1 unfortunately.

Unrelated - Is 75C the max anduril can set to?

We shouldn’t have to do any of this of course, but at least some are possible.

Ramping Problem not solved, but a workaround - for me at least…
Mine does start dropping off at the 20-30 secondish mark.
When it does that, double click back to turbo, and then it stays there for quite a while. It does seem to be a problem of speed of temperature rise, not an actual temperature problem. Since it becomes somewhat stable after initially heating up, it stops doing the yo-yo, for me at least. The external temp settles around 60C with the ambient set properly and max temp set to 75.

Anyway, this is my take on the situation. Maybe some other people have opinions on the situation?

PS, I know muggle mode steps down early on the FW3A due to some software bug. Does it do it on the FT03S as well? I don’t have much to contribute on this topic. Maybe the bug can be fixed on later revisions of Anduril. We don’t even know what version Astrolux is using in the FT03S. Might be pretty old.

It would be interesting of someone with both a FT03 and FT03S were to swap the SBT-90.2 into a FT03. Then see how that driver and reflector combo acts. How is the stepdown working? Much better, poorly, acceptable?

See how many amps it pulls, I suspect no change here.

See how the throw improves. We already know it does, but still not as far as the rated specs.

The counter argument to this is that the people who own other FT03 variants have swapped reflectors, and are still _significantly _under the advertised throw.
I seem to recall something like 280 with bad reflector, and 320 with a “good” reflector. Nowhere near 510.
So maybe they tried to redesign the reflector for the FT03S, and either did a poor job, or there was a manufacturing issue with them. Alternatively, they calculated the theoretical throw (and potentially messed that up too) of the SBT90.2 in an FT03, and slapped that number on, even though it’s completely unreachable with current manufacturing limits.

It’s anyone’s guess though. M4D MAX has reached out to some contacts to try and get some info but they’re on their CNY break, I believe.

I know you can trick the light to thinking it’s colder than it actually is. Just tell it a lower ambient temp. This effectively raises the max stepdown temp.

I’m not sure what the max temp Anduril can be set to. I’ll ask Toykeeper.

The inherit problem of lower than advertised throw and lumens is due to the 20 gauge wire. The while system appears to be power starved on turbo at least. Perhaps it was a engineering choice to limit the heat generation but there was significant jump in performance when ZozzLights stepped up to 18 gauge. He measured around 14 amps on 20, I bet he was pushing closer to 20amp on 18 gauge.

He said after doing the wire swap, he got much closer to the advertised lumens, and probably candela as well. We all know heading into this as well that the SBT90.2 could eat 20 amps easily. ANSI specs 20 gauge is rated up to around 11 amps. Therefore it really leads me to conclude two options; 20 gauge could be a cost saving move, or a thermal regulation move. I’m leaning to the former.

Only 14 amps? People measured that on the xhp50.2 version. Did Zozz post that somewhere in this thread? (I may not remember). Or did he post in a different thread? (I may not have seen it) Anyway, I’d like to read up on his results.

M4DM4X got 433.5Kcd with the old FT03 reflector which isn’t bad at all. That’s 1316m of throw which is good for a light of this size, output and price. Even with a good reflector and wires I doubt it’ll be able to hit 500 cause of the single cell which won’t able to push the SBT90.2 to it’s limits. I mean it might but just at turn on. Once they fix the issues this will be a great light. Wonder what they’re gonna do with the existing ones though…like mine.

That’s a good point. The 510kcd definitely seems too ambitious regardless of whether the reflector is perfect or not. Maybe they tried tweaking the reflector shape due to the domeless design of the SBT-90.2 and accidentally made it worse? I think it’s focal point might be slightly different and/or the height of the led might not be the same as the SST-40 and XHP50.2. I haven’t looked at the led dimensions in the data sheets. So far I can only find the first gen data sheet and it doesn’t list dimensions such as the height from the base to the top of the glass. :weary:

EDIT:
The height of the base for:
SST-40 is 0.60mm
XHP50.2 is 0.65mm
SBT 90 is 0.91mm

Dome heights vary:
SST-40 is 3.01mm
XHP50.2 is 3.50mm

The Cree data sheet shows the “optical reference” at the top of the base which is 0.65mm. I assume this is reference to the focal point in which reflectors need to be designed and shimmed to. So SST-40 and XHP50.2 are virtually identical.

Edit, I found the mechanical specs for SBT-90. I assume it’s the same for Gen 2.

So there is a difference in the focal points between the SST-40/XHP50.2 and the SBT-90. Not much, but a little.

I’m sure you can set it higher. Just haven’t tried it. Don’t see why it would make a difference to the turbo runtime though as 65 and 75 both perform the same.

Something’s not right about those readings. Maybe that’s why his numbers are so low. I’ll measure mine tomorrow. Has anyone else checked to see if they also have 20 gauge wires? Maybe they changed them after they already had made some.

My memory of the numbers was way off, my bad.

Looking back through the thread, I don’t see any mention of M4DM4X changing wires, only ZozzV6. Any chance M4D M4X didn’t do a wire swap and could push that 433 even farther? That’d be fantastic.

Okay, the video. It’s long, so I can’t remember all it’s details. Lol. I prefer pictures for easy reference.

I see he was using some average cells, LiitoKala 5000mah black and a Shockli 21700 (maybe 4000mah black?). I bet it would meet the lumen rating if it had better cells with less voltage sag at higher amps. Something like a 40T or P42A, etc… Still, 4100lm-4200lm is not far off from the 4500lm spec.

FET drivers are sensitive to the total resistance in the circuit. He got higher amps using those average batteries by reducing the resistance in the led wires. The batteries also cause resistance. Keeping the stock wires and swapping to better cells will also cause a reduction is total resistance and boost amperage and lumens.

Better cells plus bigger led wires give even bigger amperage and lumens. You have to be careful here to not damage the led. It’s quite possible Astrolux erred on the side of caution to keep people from burning up leds and blaming the manufacturer for defective products.

I haven’t done anything to mine and got 415Kcd. Numbers on my meter are about 10% higher than they should be so about 380Kcd. which is same as M4DM4X’s with the FT03S reflector.

Why would your lux meter read high?

That Sofirn 21700 the M4X used is pretty decent, but there are better cells. I bet it can be pushed further.

Well my FT03 SST40 got a bit over 250 which seems a bit high. Most people’s numbers are lower and if you look at the beamshot on wikilght.de the M3XS-UT which is rated at 250 seems more intense than the FT03. GT mini also measured around 150 as far as I can remember. For some reason I only got 76k with my L16. Apart from the L16 my numbers seem a bit higher than they should. I heard that a different sample of the same meter can produce different readings cause they’re not calibrated perfectly… unless it’s expensive I guess. Also the Uni-T 383 seems to be more a bit more sensitive to cool light.

I decided to do some not-so-calibratd runtime testing using Zak.Wilson’s ‘ceilingbounce’ app on my Android phone.

My setup is very simple, since I don’t have an integrating device like a lumens tube.

setup: Flashlight was placed in a flat surface pointing at the ceiling, in a small darkened room, and with the ‘ceilingbounce’ app recording the ‘runtime’, until I manually stopped it several minutes later.

Both the FT03S SBT90.2 and the FT03 XHP50.2 were tested separately. The 26650 batteries I purposely chose higher IR 26650s (I measure around AC IR 30mOhms for this “higher-IR” 26650, where a Liitokala Lii-50A26650 would have AC IR [using YR1030 meter] around 15-18mOhms).

FT03S SBT90.2 (calibrated the thermal sensor to room temperature approximately, which I guess is around 28 degrees Celsius here in our tropical country at this time of the year). I set the Anduril thermal threshold to 60 degrees Celsius

*(the ‘glitch’ near the 9th minute mark is due to my touching the flashlight which moved it a bit)

~

FT03 XHP50.2 5700k (didn’t adjust thermal control settings for NarsilM; I forgot to check how far from the correct room temperature is the stock temperature readout)

comment:
The FT03S lumens graph seems to mirror M4DM4X’s graph — there is a big sudden drop initially (around 30 seconds or so), but it recovers and maintains it for sometime before gradually dropping again after a few minutes.
(after the testing which I manually stopped — the flashlight was very very hot and almost untouchable for my hands)

On the FT03 XHP50.2, there is a big drop at around 3 minute mark (maybe this is the stock default setting?)
(when I manually stopped the test, the flashlight was just warm)

I wouldn’t try to discount your results. I lot of folks here, including myself, use a LX-1330B which according to Djozz’s test reads higher than the 383. As long as your measuring distances, technique and math are correct, I’d use the results it gives you. I typically use 10 meters, at 30 seconds (when possible) and move the meter around in the hotspot to find it’s highest reading. Some people dont wait 30 seconds and other people my set their meter to record the highest reading. I find both those techniques to give higher than normal results.

BTW, I also measured 148Kcd on the GT mini.