Convoy S21D w/ legendary Nichia 219b. Review & comparison w/ Nichia 519a, B35AM, FFL351A ROSY & other "tint-snob" LEDs. (Summary & measurements on P. 1) LATEST :+1:: Firefly E07X with VERY rosy 351A LED and Lume1 Driver.

cannga · 2022-10-17T16:03:08.000Z

Beam Color Comparison: Comparison of LEDs in different hosts is inherently difficult; as a result the exposure setting of pics below are not the same. I adjust so they best approximate what I see in real life; meaning very subjective.

At this level all 4 LEDS are excellent in reproducing colors. They do have different tints (rosy vs neutral) and which one you like comes down to personal preference, if you like the famous rosy/magenta tint, then only 219b 4500k would do. If you prefer a neutral tint, then it is 519a 4500k with what I see as a “golden” tint. In between those 2 bookends, there are the B35AM and GT-FC40. Some comments based on “real life” view:

Nichia 219b 4500k: That famous rosy/magenta tint. In actual use in nature scenes, both GT-FC40 4500k and (even more so) Nichia B35AM resemble this LED.

Nichia 519a 4500k: Neutral tint, “golden” color (more yellow) to my eyes.

Nichia B35AM 4500k: My nick name for B35AM: “219b with throw” IMHO. If the magenta tint of 219b 4500k is too much for you, then this should provide some relief, if only a little. B35AM still leans much more towards 219b’s magenta than the neutral tint of 519a though. There are two LEDs in my collection that look like 219b: first is E21a, and second is this one, B35AM.

Getian GT-FC40 4500k: Still looks like 219b overall, but now surely leans more towards the golden tint of 519a.

dick33 · 2022-10-17T16:35:04.000Z

phantom23:

dick33:

I have convoy with B35AM, very low lumens, weak hicri led. Not more than 1000lm on 3A (18w) and it make smoke on this. Just 700lm on 9w.

There must be something wrong with yours then because B35AM should produce about 1400-1500lm at 2,4A. There were issues with burning LEDs at 3A that’s why Simon dialed it down. Maybe yours was affected as well?

B35AM may produce 1400lm but that would be 70cri 5000k on super good mcpcb and without losses on optics etc.
Convoy’s b35am 4500 r9080 will never give you 1400lm, they will fry much earlier.

I talk about real lumens in real flashlight s21a/b with 3A driver. And it give you just 1000lm on 3A. Datasheet tells the same - 1400ma and 700lm for convoy’s bin. You may find actual tests or measure it yourself.

phantom23 · 2022-10-17T18:24:30.000Z

It will give over 1400lm:

Nichia NV4WB35AM Flashlight Modding and DIY Parts

It’s not really a thrower LED. It has larger die area than SBT90.2 and produces far less output. I think Clemence has tested that but I can’t find the results now. Anyway E21A is not a thrower LED and this one will only throw less. But when high CRI is a requirement, there are very few choices - and this is a fair one. ADDED: [image] The high CRI variant does not much more than 2000 lm at the peak. So intensity is ~3 times lower than that of Boost HX.

R9080 (high CRI), E800 (which is exactly what Convoy uses) produces around 1250lm at 2A, 1400lm at 2,25A, a touch over 1500lm at 2,5A, it should produce ~1750lm at 3A unless it fries itself. Simon reduced the drive current from 3A to 2,4A in order to avoid the latter.

We’re talking about around ~1450 emitter lumens. GT-FC40 driven at 2,5A (12V so essentially double the power) - 2600 emitter lumens. 519A at 6A (3V) - ~1250lm. Those drive currents are what Convoy uses in M21B:

aliexpress.com

26.89€ 14% OFF|MAO Convoy M21B 21700 taschenlampe| | - AliExpress

Smarter Shopping, Better Living! Aliexpress.com

dick33 · 2022-10-17T19:49:30.000Z

phantom23:

It will give over 1400lm:
Nichia NV4WB35AM - #36 by Agro

R9080 (high CRI), E800 (which is exactly what Convoy uses) produces around 1250lm at 2A, 1400lm at 2,25A, a touch over 1500lm at 2,5A, it should produce ~1750lm at 3A unless it fries itself. Simon reduced the drive current from 3A to 2,4A in order to avoid the latter.

We’re talking about around ~1450 emitter lumens. GT-FC40 driven at 2,5A (12V so essentially double the power) - 2600 emitter lumens. 519A at 6A (3V) - ~1250lm. Those drive currents are what Convoy uses in M21B:
26.89€ 14% OFF|MAO Convoy M21B 21700 taschenlampe| | - AliExpress

I know this graph. In convoy it’ll be a ‘little’ different result. But once again - if you take real convoy flashlight and measure it - no way you will get 1400lm with its b35am. You operare with theory not practice.
Measure it yourself or open testing results of s21a/b b35am.

You may not trust me and belive that u get 1500lm.

phantom23 · 2022-10-17T20:09:32.000Z

As I said those are emitter lumens, before optical losses but they’re like 15-20, not 40. Most reviews of Convoys with B35AM are pretty old when the LED was overheating. Also newer ones have different drivers where S21A pushes just 2,2A into the emitter which is less than in M21B.

But still it should produce almost 200lm more than 519A and at 2,4A it shouldn’t fry itself.

cannga · 2022-10-17T21:24:38.000Z

Thanks guys for the interesting discussion. The graph that @phantom23 was referring to (the red curve):

Nichia NV4WB35AM Flashlight Modding and DIY Parts

It’s not really a thrower LED. It has larger die area than SBT90.2 and produces far less output. I think Clemence has tested that but I can’t find the results now. Anyway E21A is not a thrower LED and this one will only throw less. But when high CRI is a requirement, there are very few choices - and this is a fair one. ADDED: [image] The high CRI variant does not much more than 2000 lm at the peak. So intensity is ~3 times lower than that of Boost HX.

phantom23 · 2022-10-17T21:32:16.000Z

Yes, those are emitter lumens before optical losses.

cannga · 2022-10-19T19:21:48.000Z

cannga:

External Temperature after 25 minute run-down from Turbo/Max brightness level, measured with IR gun pointing at the head area…

Convoy S21D —-> greater than 60° C
Convoy M21B —-> 59 C
Emisar D4V2 219b 4500k (with max thermal limit set to 55 C) —-> 49 C
Olight M2R Pro —->47 C
Sofirn SP36 BLF 4000k —->45 C
Convoy M21B w/ B35AM 4500k —->42° C

Observation from “real life” (not white wall hunting) use:

The M21B w/ B35AM is the coolest running Convoy flashlight I have, by a wide margin, clocking in at 42° C after 30 minutes continuous on from a Turbo start at home. In actual use (Turbo start then leave light alone during my walk), the S21D w/ Nichia 219b becomes almost/kinda too hot to hold in current LA weather, M21F is second, a little cooler but still quite hot, but M21B is acting as if it’s just loafing around.

Even more surprising, it is THE coolest running flashlight I have including other brands. At 42° C it acts as if it’s running at half throttle. I almost wish for a higher voltage (edit: current, thanks phantom23) than 2.4 A but I understand bad things did happen at higher current so unfortunately that won’t happen.

Allow me to post another picture of this very pretty light, together with my other latest Convoy light, M21E with Getian GT-FC40 at higher CCT 5500k. It is EXACTLY the light I have been looking for, a Nichia with throw and I recommend extremely highly (I love it).

phantom23 · 2022-10-19T19:24:33.000Z

B35AM may appear to run cool but partially that’s because of poor thermap transfer (that’s why Simon had to reduce current from 3A to 2,4A - emitters were frying).

cannga:

I almost wish for a higher voltage than 2.4 v but I understand bad things did happen at higher voltage so unfortunately that won’t happen.

Of course you meant higher current.

cannga · 2022-10-19T19:30:00.000Z

phantom23:

B35AM may appear to run cool but partially that’s because of poor thermap transfer (that’s why Simon had to reduce current from 3A to 2,4A - emitters were frying).

cannga:

I almost wish for a higher voltage than 2.4 v but I understand bad things did happen at higher voltage so unfortunately that won’t happen.

Of course you meant higher current.

Yes of course thanks, current. :+1:

Questions please: Why would it have poorer thermal transfer than say GT-FC40 in same M21B host (ie same design)? My GT-FC40 in M21B stays at 60° C.

Also this is a 30 minute run so the thermal transfer can’t be that poor, can it? Wouldn’t it fry the electronics if all that heat does not escape after 30 min?

phantom23 · 2022-10-19T20:39:42.000Z

Poor heat transfer between the emitter and PCB. There are no dedicated PCBs for B35AM, flashlight manufacturers need to improvise but thtere’s no good solution as of yet. So yeah, the emitter itself is hot it just transfers less heat to the body. That’s one thing. At 2,4A it uses 50% less energy than GT-FC40 so it has to run cooler.

jon_slider · 2022-10-19T22:24:48.000Z

cannga:

The M21B w/ B35AM is the coolest running Convoy flashlight I have
…
the S21D w/ Nichia 219b
…
M21F is second
…
M21B…

If you determine the Actual Sustained Lumen output of each light, and its weight, you can calculate lumens per gram.

the coolest running light probably has the lowest Lumens per Gram.

QReciprocity42 · 2022-10-20T05:59:22.000Z

@cannga: what is the ambient room temperature when you did the tests? The equilibrium temperature of a running flashlight depends heavily on both amount of power dissipated by the light AND the ambient temperature--leaving out the latter makes it impossible for us to deduce how much an identical light would heat up in our usage, as our ambient temp might differ much from yours.

cannga · 2022-10-20T19:04:52.000Z

QReciprocity42:

@cannga: what is the ambient room temperature when you did the tests? The equilibrium temperature of a running flashlight depends heavily on both amount of power dissipated by the light AND the ambient temperature—leaving out the latter makes it impossible for us to deduce how much an identical light would heat up in our usage, as our ambient temp might differ much from yours.

It’s good you’re taking an interest in this. Nothing unusual 70-75 F. But let me rerun the test with all 4 lights (M21B with B35AM and FC40, M21E and M21F with FC40) in same condition tonight just to make sure I haven’t screwed up somewhere.

Are you thinking convection heat transfer also as a method to check what the temp inside is? Would shooting at the lens provide a clue also for internal temp? I will also take the lens off and check the LED, but it might too hot (upper limit of my IR gun is 60 C).

cannga · 2022-10-20T19:05:06.000Z

From Simon 4/17/2022
I’m sorry about the smoke from the B35AM. I have tried to find MCPCB with high thermal conductivity, but without success, the thermal conductivity of the MCPCB insulation layer currently used is 1.0.
I adjusted the output current of the driver to 2.4A and the LED no longer smoked.
I will be producing a batch of 22mm B35AM custom drivers with an output current of 2.3A.

Interestingly enough, the smoking was observed AFAIK in S21a (anyone pls correct me as needed), a smaller light than M21B. That combination of S21a and B35AM was released around 1/2022. After the above post in 4/2022, the current was reduced to 2.4A.

My M21B version was released AFTER the reduction, so it never went through the “Armageddon challenge.” You think it might just tolerate the higher current since it has a larger thermal mass? From actual use, seems like this M21B is just loafing during a Turbo run-down, with excellent sustained brightness.

QReciprocity42 · 2022-10-21T02:40:13.000Z

cannga:

It’s good you’re taking an interest in this. Nothing unusual 70-75 F. But let me rerun the test with all 4 lights (M21B with B35AM and FC40, M21E and M21F with FC40) in same condition tonight just to make sure I haven’t screwed up somewhere. Are you thinking convection heat transfer also as a method to check what the temp inside is? Would shooting at the lens provide a clue also for internal temp? I will also take the lens off and check the LED, but it might too hot (upper limit of my IR gun is 60 C).

Thank you for the reference! I am very interested in seeing how the re-run turns out. My room temp is probably hotter than yours, and my S2+ 519A triple running only 700mA mode gets noticeably hotter than body temp after half an hour. On the other hand, it's below 45F outside, and my C8 running SFT40 on ramping driver can sustain turbo indefinitely.

I have no idea if convection heat transfer works. I don't have any knowledge or experience using an IR gun. But my guess would be to avoid reflective surfaces like the lens. I would be very curious to see your LED temp measurements!

cannga · 2022-10-21T14:42:08.000Z

Convoy lights were re-tested @ room temp 26 C. I tail-stand the lights and start at Turbo level, then measure temp in next 30 minutes by pointing IR gun at the flashlights’ head, about 1 cm away.

The S21D with 519a immediately heated up to 60 C (limit of my IR gun), within ~5 minutes. The other lights (M21B x 2, M21F, and M21E) reached equilibrium at around 20-25 minutes. Temps at the end, 30 min after Turbo start:
M21B w/ B35AM: 42 C
M21E w/ FC40: 44 C
M21F w/ FC40: 50 C
M21B w/ FC40: 57 C

Glad I did the test; didn’t realize that there is such a large temp difference between the larger M21E versus M21B. I have all 3 so it doesn’t matter but if I have to pick one out of these 3, for me this is another reason to pick M21E over the other 2 FC40 lights.

Holding the flashlights with cooler temp is obviously more comfortable. S21D with 519a too hot to handle inside house, and better outdoor.

I tried to measure at the lens and also the LED directly after taking lens off, but it was too hot, beyond limit of the IR gun (60 C).

Caleb · 2022-10-21T17:10:43.000Z

cannga:

Convoy lights were re-tested @ room temp 26 C. I tail-stand the lights and start at Turbo level, then measure temp in next 30 minutes by pointing IR gun at the flashlights’ head, about 1 cm away.

The S21D with 519a immediately heated up to 60 C (limit of my IR gun), within ~5 minutes. The other lights (M21B x 2, M21F, and M21E) reached equilibrium at around 20-25 minutes. Temps at the end, 30 min after Turbo start:
M21B w/ B35AM: 42 C
M21E w/ FC40: 44 C
M21F w/ FC40: 50 C
M21B w/ FC40: 57 C

Glad I did the test; didn’t realize that there is such a large temp difference between the larger M21E versus M21B. I have all 3 so it doesn’t matter but if I have to pick one out of these 3, for me this is another reason to pick M21E over the other 2 FC40 lights.

Holding the flashlights with cooler temp is obviously more comfortable. S21D with 519a too hot to handle inside house, and better outdoor.

I tried to measure at the lens and also the LED directly after taking lens off, but it was too hot, beyond limit of the IR gun (60 C).

Cannga,

Thanks for doing this retest. I appreciate all the time that you take to provide useful information (beamshots, testing, insight, etc).

You add a great deal of value to BLF.

Cheers!

cannga · 2022-10-21T22:22:12.000Z

@Caleb thanks for the kind words. I’ve learned A LOT from people on BLF also - MANY souls here with “scary” deep LED and flashlight knowledge.

It is so interesting that the Convoy lights, once they hit auto regulation limit of 55° C, seem (ie speculation) to go down to a certain brightness level (~37% of max the few times I’ve measured w/ ceiling bounce) and stay there, almost seemingly regardless of the resulting temperature. For example S21D settles at 60 C, whereas M21E goes to 44, both with brightness ~37% of Turbo start.

What I find interesting is that although Tmax is 55 C, S21D steps down but then stays at 60 C externally. (Not complaining at all. I am a HUGE Convoy fan for this reason - it brings uniformly high sustained brightness.)

cannga · 2022-10-21T22:35:07.000Z

Next I would like to do the same temp measurement as above except starting not at Turbo (100), but at the next lower power level, 35 for rear switch, and 40% for e-switch Convoys.

I know it doesn’t make sense but sometimes I feel like (ie not sure) +starting at the lower level 35+ results in a cooler light than starting at Turbo and then let it stepdown to 37. Even after 20 minutes for equilibrium. Not sure about this though, it’s just subjective tactile eval so probably the test will prove me wrong.