The BIG LED SFH55 - 11 x 10 mm 3V no dome monster

LED Flashlights – General Info
1

I received a sample of this LED from Neal - no info, no labeling, no docs. He just asked me if I'd be interested in testing a new LED. I couldn't resist. After some simple bench power supply tests and measurements, I realized it's a 3V LED, must have a pretty low Vf so should be easy to get some serious amps feeding it, and it's footprint matches a SBT90.2 so we should have available MCPCB's to fit it. I decided to host it in a WildTrail WT90 since I have 2 of these, should be easy/quick to re-use the same MCPCB, and has 3 21700's to power it.

I've posted most of this info in the thread Dave started on these high powered, new generation Chinese LED's: https://budgetlightforum.com/t/-/67698, but realized I scattered the pics and info in way too many posts, and it would be better tightening it all up in one post, dedicating a thread to featuring just this LED, as it deserves, so here it goes...

Initial closeups, comparing to other large known LED's: XHP70.2, MT-G2, then SFH55 showing the 4 x 4 grid:

16 LED's in one?

Same dims but the dye surface sure looks tiny on that SBT90.2:

This is the light I'm using, shown here for reference:

Reflowed on the WT90 MCPCB. The LED+ pad on the LED has a notched out corner which matched the MCPCB. Funny because the SBT90.2 had the notch on the opposite end of the pad, so didn't match the MCPCB (not shown here though):

Single pair of 18 AWG wires:

Nice moon mode with Anduril 2:

Ceiling shots of the beam. These match pretty well to what I'm seeing. I'm not seeing much, if any, tint shift, hot spot is pretty well defined:

Wall shots. Not happy with these but it shows the hot spot in the size I can see with the halo around it. The spill is definitely darker in the pics, and at this distance, the center of the hot spot actually has a yellowish tint which I could not get to show in the pics:

Measurements

After reflow on the WT90 MCPCB, first test:

  • On 3 30T’s at 4.20V in the carrier: 10302 lumens at turn ON, at 2 sec pause: 514 kcd
  • On a single 30T at 4.20V, using a 14 AWG wire on the tail: 35 amps, 9880 lumens at start

On a set of 3 Lishen LR2170LA cells at 4.20V:

  • at start: 10390, at 10 secs: 9450, at 30 secs: 9030
  • throw: 511 kcd, taken at 5 m (2 sec pause)

On a set of 3 old gen 40T's at 4.20V:

  • at start: 10390, at 10 secs: 9420, at 30 secs: 8910
  • throw: 511 kcd, taken at 5 m (2 sec pause)

Impressions

The beam has no XHP tint shift halo effect, but did notice in closer distances, say under 10 feet, the hot spot is consistent white but further out, still indoors, the hot spot has an inner more yellow spot and the hot spot itself gets fuzzy around the edges. Focus seems ok because you only see a center black spot under 1 ft or so. But dunno - didn't play with focusing at all.

Overall the tint isn't bad for a cool white, I'd guess about 6000K, maybe slightly lower. The spill tint is very close to a SBT90.2. The lumens suggest the ratio of amps to output is in the ballpark you would get from a triple XHP50.2 3V light, so that's decent for a domeless LED. The candela is about half of an SBT90.2 in the same light, but the lumens is about 80% higher while drawing about 15-20% more amps.

WT90 w/SBT90.2 test results on 40T's at 4.20V:

lumens at start: 6330, at 30s: 5180, throw: 1055 kcd (estimate about 27-28 amps).

This LED can reportedly do 9100 lumens at 25 A and 14200 lumens at 62 A. from my tests, the 9100 at 25 A is a little optimistic, but for the higher end of 14200, with current driver technology and availability, I don't see how we could do much better than 35 amps. On possibly 6-8 21700's in parallel, maybe we could break 40 amps but that's about it.

I was more optimistic about this LED earlier in testing, but after reviewing these results, I'd say I got mixed feelings about it. It's slightly better in lumens output than a single XHP70.2 at it's best, but the throw is significantly less than a SBT90.2 at it's best, so not sure it's really worth it, though I am more slanted towards throw capabilities, and I do expect this to be a lot cheaper than a SBT90.2, maybe less than a XHP70.2 as well. Hopefully we'll see it in a few lights though because of it's cheaper cost.

I also haven't seen it much outdoors at distance yet. The indications I'm getting is the hot spot tends to blur out on the edges at greater distances so I'll look at this closer soon, but could be very dependent on the reflector.

2

Interesting LED to say the least.

Like a lot of LED’s wish it had warmer tints and high CRI options, without that can’t think of any uses for me but still good to see other options hitting the market.

3

I agree.

Tom E thank you so much for testing this and posting your results.

4

+1.

BTW, am I the only one thinking about GT Nano? :smiling_imp:

As to the AMP limit, I guess we need some high power buck drivers to really get the most out of it. Lexel did some for SBT90.2 but even that is not powerful enough.

5

4S NiMH and low resistance or 2S LFP with higher resistance could be a low-tech way to get the amps

6

Because somebody told me he was working on a high CRI 3mm round LES 3V LED that can take up to 20A I looked at how to drive it with DC-DC converters from 1S or 2-4S, buck-boost should be possible up to 25A (as long as there is enough board space) , and buck up to 40A with the LM27403 from TI. I started looking at components to use but then he told me he abandoned because the high CRI phosphors cost way too much and so I lost interest.

Anyhow I think it’s doable to make a high power buck driver for these fat LEDs (after all Acebeam use buck drivers in their high power lights AFAIK), and if they were high CRI I would work on it.

LTO is around 2.2~2.3V nominal, closer than LFP in 2S, but it has pretty low capacity, can be severely abused though.

7

Than you for presenting your analysis of this exciting LED.

I can think of five dedicated thrower flashlights that I would like to have this LED in: Small-medium dedicated thrower, medium size, large-medium size, large and extremely large dedicated thrower.

Same thing with power throwers and power flooders, too. That's 15 new flashlights right there. But I would want Mateminco MT90 Plus SHF55, Lumintop BLF GT 4x SHF55, and Imalent MS18 SHF55, first, then all of the others.

8

Rich, I like the way you think! The MS18 with 18 of these monsters -- hhhmmmm.

9

Lol. Speaking about this light, zeroair’s review shows it’s unregulated with what looks like PWM dimming. It seems like they are direct driving those XHP70.2, either 2S to 6V or 4S to 12V.

10

Lexel used multiple driver ICs in parallel to increase the max AMPs. He basically took one of his regular designs and multiplied components. And only in a later iteration he did some optimizations.

11

What’s the price of them compared to the SBT90.2? If it’s a lot cheaper then that could be a good reason for some people

12

Yep, LM3409, rated for 5A max but I think Lexel used it up to 8A , it’s a LED driver so it’s easy to implement compared to general purpose converters, the downside is that it’s asynchronous hence less efficient, and to reach high currents with multiple ones a lot of board space is needed.

13

Pure orgasm !!!

14

Info is sketchy - we don't know where to even buy them, but from the chart in the OP here, it implies the price is just under 1/3 the cost of a SBT90.2, speculating in the range of $10-$12.

15

Have you measured the output of the FW21 Pro XHP50.2 or E07 XPL HI/SST20? Wonder how it compares. It seems like this emitter would be a good alternative to 3x XHP50.2 but with increased throw.

16

This is a good point. In the example case of FW21 Pro 50.2, by going to a single emitter / reflector, you’d gain ~65% optic area. On an emitter surface area basis, this is about equivalent to three shaved 50.2s. ZeroAir reported ~9200 turn on lumens with a 30T. TomE states 9880lm turn on with a single 30T. So some napkin math, 1.07x for lumens, 2x for apparent emitter area (for domed 50.2s), and 1.65x for the single reflector… That’d be ~3.5x the candela with similar lumens. I like it.

17

I got #'s on a triple XHP50.2 modded Amutorch X9 - don't have a FW21 Pro, not a fan of any tail e-switch lights actually. I do have an E07, think SST-20's. I'll check the #'s.

18

How is the tint compared to the lights mentioned above? Is the hotspot cooler than the XPH50.2?

Not a huge fan of the tail e-switch either. I regret selling my E07…I do like some things about the FW21 pro but it has it flaws. It used to be about 30% brighter than my E07 at turn on but now it’s almost the same. Maybe I need a new battery? I also noticed that the turbo output can change just from loosening and re tightening the head. On the plus side I do prefer the more efficient LEDs and constant current driver, but then again, it lacks aux/switch LEDs and doesn’t have the best tint. I’d like to see a light the Size of the E07 but with this LED. Should be super bright but with decent throw.

19

I have a FW21 Pro and thinking of something like that the one that would be a good interesting beast is a Pioneer K36 with SFH55. What do you think?

20

Found where to buy the LED, Pioneman store on aliexpress sells it for $13 on a 24mm pcb. Might be interesting to put in a Convoy 3x21A with triple molicels, if the convoy driver can make use of it. Hopefully over 10000-11000 lumens if the tailsprings are bypassed.

https://www.aliexpress.com/item/1005003890697154.html