show of hands for a new Lux3 driver design?

[Leeoniya]

i've been doing some research on constant current and DC/DC converter step-up ICs.

my design specs required as low input voltage as possible, small surface mount package, high efficency - 90% or more, and reasonable current handling capacity...with enough overhead for small error to drive a lux 3

i came upon the LTC3426 and decided to see if i can find it anywhere on the CPF forums as being mentioned in a single cell or multi-cell design. the search yielded no matches over the past year. so i am now set on designing a drop in module based on this IC...it has the potential to be very small and very efficient since it's SOT-23!!! and has few components other than the inductor that need to be large...which is also relative.

datasheet

it will most likely be similar to a MadMax...i dont have a MM so i dont know what the MM design is. but i've never seen the use of an sot-23 IC in a drive application before so i think it would be intriguing to make one. and i have a feeling i can make it quite small.

maybe if there's enough interest i might make many and sell them for something reasonable. or i'd be happy to split the revenue with anyone who has good connections for mass producing surface mount stuff.

either way, i'll post schematics and a parts list.

i'll also try to make it the circuitry as tight as possibly to allow intyegration into the smallest diameters of flashlight, and try to make it very easy to swap out emitters...eg having the solder pads free of nearby components.

Leon.

 

[andrewwynn]

the tiny driver ICs ive seen were from TI.. i have a couple but haven't had time to build them into lights.. the nice thing about the TI models.. more efficient at lower power levels, opoosite of all the other guys.. it's about 85% efficent at 400mA, but 95% efficient at 70mA.

I ended up building an LDO driver for my project, so i didn't need to build that one yet, however, since it's so much more efficient than the LDO at low power i'll be revisiting it...

my driver board is 10mm diameter to fit the head of an arc/peak AAA light.. i'd love to see what you come up with.

-awr

 

[idleprocess]

Sounds interesting. I've had project simmering for a while now that will require a decent 3W driver. I was thinking of powering it with 6V, but if there's a nice 3V driver that can deal with the discharge curve for 2xNiMH, that's even better.

 

[NewBie]

Well, some comments to help you along.

Decent output current is considered these days 0.7-1A.

Better than 90% efficiency as you said.

At these current levels, it is important to have a constant current output, as you can set the voltage one day with a new emitter, and the Luxeon LED Vf drops significantly over time (an entire Vf bin drop) and with temperature.

Consider a Surefire or Eveready 123 cell driving a converter at these type of levels will drop to 2.54 Vout fresh, and head south from there.

Alot of folks would like a converter that works good off a Li-Ion, 4.2V-2.5V discharge, pushing a 3.9V LED at 1A. This requires a buck-boost, a hybrid, or a sepic.

What most folks forget is that many cells sag quite significantly under load, such that a 123 is pushing 1.7A and 2.54V when a 90% eff. converter is driving 3.9V at 1A.

Then the battery sags on down to 1.8V over time, requiring the 123 cell to deliver 2.4A to hold 3.9V on the output at 1A.

The 2x NiMH AA situation is worse.

 

[HarryN]

As Newbie pointed out, it really needs to be a buck / boost. The common applications expected would be
- 1 x AA NiMH nominal 1.2 V or 2 x this cell
- 1 x 123 Li nominal 2.2 - 4 V or 2 or 3 of the same
- Lux III or Lux V
- 3 drive current settings, maybe 25, 300, 1000 ma

I would not need it, but some would argue correctly that any setup that you want to really sell needs a battery level indicator, and the ability to deal with avoiding over discharge.

If you want to sell it to people that live in cold climates, it needs to use chips designed (not assumed) to run at (-40 C). For our dessert friends, which is where a lot of donated lights end up (see the group in Iraq the club supports), it routinely hits 120 F temps there, + the heat of the LED.

SOT 23 is small when I have to work with it, but it is not so small when you have to cram it in a light. Last but not least, it takes some real thought to pot the package and indutor to keep them from shock destruction.

OK, not that I have made your life difficult, feel free to work on it and make mine easier. /ubbthreads/images/graemlins/smile.gif

 

[LED_ASAP]

Forget about the 3426 chip. The 0.11ohm internal switch alone will fail your 90% efficiency requirement at 0.7-1A output level. Please remember that the 1A is the output (I assume you mean average current, and not peak current that some manufacturers seem to claim), and the peak input current will likely be close to the 3A range. The average battery draw, as NewBie pointed out, is likely in the 2A range. If you want a 90% efficiency converter, you really have to limit your total resistive loss to less than 0.1ohm---and that is including the Ron, the ESR of the inductor, and any ground-path resistance. It is not easily achieved even with new chips these days.