It's been years .... What's going on ;)

[kasey197]

For one reason or another it's been a long time since I've checked in...

Would you believe my kid is now into my old hobby
He wants to mod one of my old lights - single 18650 direct drive - with a brand new emitter
He says you can push more than 700lm out of a single one these days?

Is that for real ? What's the current king-of-da-hill led emitter for plain old dazzling bright at 3.7v ?

who sells me these days ? I see aw batteries are still rockin but who sells emitters - there was someone from Australia I think ?

glad to be back and happy the forums going strong!

 

[RoGuE_StreaK]

Yeah 700lm is easy, I haven't looked at any direct drive stuff so don't know what the initial current would/could be, but to give you an idea the probable "king-of-da-hill", the Cree XM-L2 (price-vs-bang-vs-availability-vs...) spits out about 1000lm at 3Amps, and has been pushed over 6Amps by some...

You are probably thinking Cutter in Australia, but personally I've found their shipping to be exorbitant; haven't checked in a while though, maybe they've changed. But I can usually find what I'm after, cheaply, from the likes of LedRise, Illumn (formerly Illumination Supply), FastTech, even BangGood or DealeXtreme for generic-binned stuff.

Something you might want to look into, and look for, is the mounting board; some are a lot better than others as far as removing heat are concerned, which when you are pushing the emitter to the extremes (direct drive etc) is going to be a major concern. Direct-to-copper is "da shiz"

 

[Anders Hoveland]

Would you believe my kid is now into my old hobby
He wants to mod one of my old lights - single 18650 direct drive - with a brand new emitter

It's like I've been saying- many of these things are genetic.
If you just grabbed an average kid from another house, what do you think the chances are you could get him interested in this, no matter how hard you tried?

 

[StarHalo]

(..but if there's space for 2xRCR123, 2000 lumens is entirely possible via MT-G2)

 

[DIWdiver]

The top bin CBT-140 from Luminus is minimum 3680 lm at 21A. Should be over 4400 at 28A.

Sure, it's not competitive in efficacy or economy, but OP asked for "plain old dazzling bright at 3.7v".

In the US, Mouser and Digikey stock pretty well. I've been buying from LEDSupply. DealExtreme and KaiDomain ship worldwide for cheap, but their stock is spotty and specs questionable.

 

[kasey197]

Thanks folks!
Having worked with semiconductors for different reasons I'd sorta always looked at LEDs as just another kinda semi - all semis hate heat and if you can't get the Thermal resistance down then that's just it - you ain't having more brightness.
I hadn't imagined that they figured how to get the thermal resistance down so much !
The Cree xr-e was about 8 degC/W. I just saw that the XM L2 is 2.5 !!
Now that just about makes a muscular 4A drive at around 3.7V very possible without cooking the junction of love

I just learnt now about the Luminus - I had never heard of them so very many thanks for that DIWdiver - I don't know about 20 amp drives but some googling shows that the Sst-50 parts looks available from kaidomain and deal extreme (thanks for the suggestions !) on star pcbs but not avail at the highbrow stores.

Does anyone have any views on these sst50s vs the Cree xm L2 for a directdrive project ?

 

[sven_m]

Does anyone have any views on these sst50s vs the Cree xm L2 for a directdrive project ?

The SST50 is much older and very inefficient in comparison.

I measured luminance of the SST50 (unknown flux bin, 5mm² die) and the XM-L2 (U2, ~4mm² die),
at 5 amps, both without dome, on copper stars. They compared like 45 : 115.
I haven't taken lumens, but the ratio is the same, as both were dedomed.

 

[scout24]

Oveready direct drives the XM-L2 in an M2 head with a single protected 18650 for north of 1000lm out the front, with nice runtime. Single mode goodness. IIRC, somewhere near 5+ amps, McClicky is no good so a zero resistance no spring tailcap is recommended. The XM-L2 seems no worse for wear with good heatsinking, and the cells live long happy lives at 2c and lower discharge. This setup has good output on a single 123 primary as well. (100lm+) The thread is worth a read, over in the Oveready subforum, Custom Builders and Modders. Enjoy!

 

[kasey197]

Thank you - these are great leads
Sven , I came to a similar conclusion as you did after i spent some time poring over the datasheets - I've ordered some (exxpen$$ive !) SST-90s on a star and will like to do a shootout with an XM-L2 and see where we end up. I cannot find the CREE's ready mounted on a star MPCB at the usual online suppliers like digikey though
Any leads on where I might be able to buy some good bin XM-L2 parts ready mounted on a star?
DX etc are just a plain crapshoot as far as bins are concerned I reckon.

I just checked out overready's DD lights thread as suggested by scout - a REALLy great read!

 

[DIWdiver]

Luminus doesn't even show the SST-50 in their product lineup any more, and there's very limited stock available. It was never a great part. Lots of lumens, but poor efficacy and low luminosity made it a dog from the get-go, at least for most applications.

Lots of lumens per die seems to be the only thing that Luminus has going for it, or ever did. Round emitter areas on some products might be aimed at theatrical applications, but Cree parts masked to round outputs can still kill them in efficacy if not in output. Their single-die white product line is narrowed to xxT-90 and CBT-140. They must be surviving on other product lines. I don't think Luminus is going to be a player in the flashlight field for the foreseeable future.

 

[DIWdiver]

Any leads on where I might be able to buy some good bin XM-L2 parts ready mounted on a star?
DX etc are just a plain crapshoot as far as bins are concerned I reckon.

Try LEDsupply.com. I've had good luck with them.

I agree DX is questionable, but I've not complained about my results from them either. Of course, I've never verified the product they sent me.

 

[kasey197]

Ok perfect just put in an order and now sit on my hands for a while until they get here.

Lots of lumens per die seems to be the only thing that Luminus has going for it, or ever did. Round emitter areas on some products might be aimed at theatrical applications, but Cree parts masked to round outputs can still kill them in efficacy if not in output. Their single-die white product line is narrowed to xxT-90 and CBT-140. They must be surviving on other product lines. I don't think Luminus is going to be a player in the flashlight field for the foreseeable future.

I dont know if thats entirely it though ? The thermal design of the SST-90s is pretty impressive - and ultimately if you want to drive a LED hard, you need to figure the heat generation. If we assume we want to live with the same max junction temperature for both LEDs - their lower heat resistance design means that the SSTs can be driven at 3.5x the current of XM-L2. Based on your comments though, it seems likely that real life considerations of max current draw (aka run times!), efficiency etc are enough to tilt things away from the SSTs (at least for flashlights) and probably the killer blow is the crazy prices of the SSTs !

Ultimately only an experiment will tell me for sure how things work in real life so thats the way I'm going to go!

 

[Fireclaw18]

If you really to have max lumens out of a single 18650 your best bet is to go with triple XPL on a 20mm star with a Carclo triple-TIR optic. Combine it with the BLF17DD FET driver and a quality IMR 18650 cell and you get a small-tube shaped single-cell light with output approaching 4,000 lumens.... 50+ watts of power. It's like having 4 car headlights in your pocket!

The FET driver gives direct-drive performance at max power, but has the advantage of also allowing your choice of lower power modes, turbo timer for safety at max power, and off-time memory for much more intuitive mode switches. Downside of a FET-Triple XPL light is you can't leave it on for long at full power as the head will get HOT. Battery will also run down quickly. This might not be the best choice to introduce your kid to modding flashlights.

If you want something much more basic, just get an XML2 emitter and a standard 3 amp driver. Should easily produce over 900 lumens with reasonable runtime, and won't be a fire hazard like the triple-XPL/FET setup.

If you're interested, Mountain Electronics has all the parts needed to make a light of either type.

 

[Fireclaw18]

(..but if there's space for 2xRCR123, 2000 lumens is entirely possible via MT-G2)

Triple-XPL, FET driver, and an IMR 18650 should pretty easily beat the output of an MTG2 powered by double IMR 18350. The XPL setup being single-cell is also much safer than a 2-cell in series setup. No risk that one cell will deplete before the other, reverse charge and then explode.

 

[kasey197]

Luminus SST-90 Direct Drive

Ok so fwiw, I tried the SST-90 on a star (bought from Digikey) and replaced the 20 mm emmitter in a generic single 18650 UF torch.
In the stock setup, with a IMR battery, the torch fires up ok, temperatures stayed below 50 degrees C but brightness, while a dramatic step up from the old XRE in this light, was really nothing to write home about. The resistive losses in the tail switch were very significant. In fact, these low volt, high amp setups are just a PITA when it comes to passive resistive losses - a small 0.02Ohm resistance is enough to cause 0.1V of drop at 5 Amps. When you look at the Vf vs I charts its obvious that this is a big deal....

I suppose there's a reason why Edison lost and Westinghouse won

So, in the interest of science (and my insance curiosity) I just popped a small copper shim into the tailcap and screwed it on - wooo hoo !! insane bright light!!

I switched it off after 15 secs and measured the emmitter temp - an unbelivable 140degrees. Theres just no amount of heatsinking that will make a pure direct drive workable for a SST-90 ... unless you start throwing in imperfections to rob the drive and if you do that then whats the point anyway

So we press on now with the Crees - lets see what joy that brings us

 

[DIWdiver]

Re: Luminus SST-90 Direct Drive

I suppose there's a reason why Edison lost and Westinghouse won

Yep. Low voltage and efficient power distribution are like oil and water.

 

[DIWdiver]

Re: Luminus SST-90 Direct Drive

Theres just no amount of heatsinking that will make a pure direct drive workable for a SST-90 ...

Well, no reasonable amount....

Actually, I'm surprised that an IMR has enough voltage to drive an SST to high power. Maybe I need to take a look at the new SST data sheets. It's been a few years.

By the way, it is possible today to build a regulated driver that would have only about 0.1V dropout at 9A, so maybe the fact that DD sucks isn't going to be a problem much longer.