UltraFire SK-98 XM-L2 U2 (FT's SKU4521200)

UltraFire SK98 XM-L2 U2 from FastTech (SKU4521200)

A couple of pics:

Zero info about the driver (8-pin multichannel MOSFET?).

Does that emitter looks genuine? Sorry but I have neither zoom lens nor decent photo-taking device.

Cheers ^:)

P.S.(2016/05/08): shrinked the pics a bit. :-)

Yup, that emitter looks genuine. Simple driver, probably not the constant current one.

Good news then, seems I did well in gambling an extra buck into this new unit. Lots of people were/are reporting Lattice Wreck emitters on the (not so much) cheaper clones.

To be honest, I had never seen an XM-L2 before but, after switching it on for the first time it's light output was certainly reeking of my good old T6 XM-L. :-)

Pill is semi-solid, certainly gapped but more solid than hollow.

A quick tailcap measurement with standard 1m AWG18 test probes on my multimeter has shown nearly 2A of draw; no load battery voltage was 3'94-3'95V (edited: tailcap current draw increased a bit with a slightly more charged battery, 2+A). Pill's thread had received some thermal paste as lubricant and heat transfer enhancement.

The torch gets uniformly warm after a few minutes of use, it can certainly be operated continuously on high without heat issues.

After checking its innards I've applied thermal paste, upgraded the led's driving wires and slighly braided and re-centered its spring.

My friend's gonna like it.

Cheers ^:)

That driver looks to be from the same family as this one. (my emitter was a Lattice Bright :confounded:
How can such a simple driver do so much?

I pulled this driver out of an UltraFire C8.
It has 5 modes, High, Med, Low, Strobe, SOS.
It does this with 1 three legged IC, 1cap and 5 resistors.
I was just wondering if somebody could explain how this very simple driver does so much.

I did a few searches on the HG5K1 IC but only found somebody selling parts for a long range cordless phone :person_facepalming:

Voodoo !
Move that - lead to the other pad and go direct drive, no charge

What have I got to lose? Its just a Lattice Bright on a hollow pill. In fact that emitter deserves no better :wink:

Yeah, they’re pretty standard in what I call the recent race-to-the-bottom Clone Wars.

The most common are the YN-20-5 and YN-20-3. My ’98 has the –5 as well.

Pretty simple. All the resistors on the bottom are parallelled 1Ωs, in this case a 0.25Ω ballast for the LED. Similar to DD but with a current-limiting resistor. The 2 different models have 4- and 6-resistors, if I recall right.

The 3-legged donk is the IC controller. There’s an R and C in parallel across the mains to “hold up” the supply volts to the chip long enough to sense half-clicks, else it starts from 0 and “reboots” it the press is longer than the RC can hold it up. It probably senses the voltage drop before it craps out totally, and that registers as a “half-press”

The output is the 3rd leg, direct to the LED through the ballast resistors.

That’s pretty much it.

Another reason why I roll my own


Oh yeah
 try getting a drop-in that’s rated 3-18V, even with a sticker attesting to that slapped on the side, and try using 2 18350s to put that to the test, and having the LED almost go incandescent.

ZERO regulation, just a small-value resistor being the only limitation when hit with twice the voltage. The LED went very blue and very dim. Miraculously survived the ~1sec it got hit with that before I broke the connection.

Nothing inherently wrong with LB emitters; they’re just a generation behind, for when “good enough” is good enough. Ie, in cheap clone flashlights.

Being that they’re often grossly undersized (eg, got an XP-E sized emitter in a cheap C8 clone), I doubt it’d survive DDing.

I’d bypass the IC and leave the parallelled resistors in series with the LED (ie, turning it into a 1-mode light stuck on “high”).


Understatement detected


While I can live with a lower output/efficiency, the main problem is that those emitters are PoSes with regards to colour rendering, maybe because whoever makes them doesn't have the know-how to make good recipes of phosphor coatings.

The bad thing is they make 'em to resemble other manufacturer's emitters, “built for faking”, instead of building their own good prestige. Screw 'em.

Cheers ^:)

I have this one.
All you have to do in this light is to order the cheapest one.
Then order some low vf emitter you like, I have the XPL V5 Warm white (cold dedome). (Nichia 219c, or xpg3 is also very good, and cheaper)
Then move the wire from the driver to the other pad.
And swap the emitter.
All under 10-15$ and very little work.

P.S. The pill is hollow, but that is why you don’t bridge any springs, and don’t swap the wires to the led. It has natural limiting. And pulls ~2.5-4A on my XPL with a salvaged laptop cell.

Yes.

Well, how is the pill on the cheapest ones? Because in the OP one it isn't bad at all, it has a hole in the center but it offers much more contact surface than those pipe-made pills, and the seating surface is thick; it certainly has good thermal transfer properties. This matters, along with receiving a genuine XM-L2.

Cheers ^:)

Like I said, a race-to-the-bottom


As long as someone understands that, it’s okay.

There’s the XM-L2, the XP-L, the XP-G3, brightness bins in the Us and Vs becoming the norm, and you’ll still see “XML-T6” ground into the side of these clones.

Those of us in the know, know what to expect, ie, the cheapest crap available that only nominally “works” for a limited time. We know not to expect the latest-gen Crees.

Those who are truly clueless and who believe that an 80-100lm light powered by 3 AAAs is “1300lm!!!111oneoneone”, aren’t about to be fooled by a lookalike LB posing as a Cree.

Hell, all the cheapo lights I bought strictly for use as hosts, have all had at least “LB” on the stars/discs, if not actual “LatticeBright”. They’ve gotten quite unabashed about it. Honestly, I think you can stop worrying about comparing emitters under a microscope to look for square holes vs round holes and trying to discern LB from Cree. You’ll probably see LB plastered all over it.

Alas, I think I’m done even with that, because the hosts aren’t as nice as they used to be. Never ever got plastic reflectors ’til very recently in some really trashy C8 clones. They’re okay for unstressed 1A XP-E2s and “bead” type 3W LEDs, but that’s it. The included switch probably wouldn’t even handle 1A for very long, let alone 2A or 3A.

Nope, once I burn through these, I’m done.

Its only on the sides. But it is not bad with a thermal paste I have used it on my bike for hours on end with 2.0-3.0A with no problems, on a laptop pull and XPL DD.

Yestedray I swapped a Nichia 219b 4000k in it. It has lower pull ~2.6A on a 4.2V battery. The light is amazing. I really cant wait for the high power 90+cri emitters to hit the shelves.